{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.12.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"none","dataSources":[],"dockerImageVersionId":28755,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# 先做基础聚合, 被迫加字段再加","metadata":{}},{"cell_type":"markdown","source":"# （1）base表：\n\n（0）聚合的唯一触发条件：规避笛卡尔积爆炸 ⭐️⭐️⭐️ \n\n> 当一张表里,同一个 case_id 对应多行数据时,才需要聚合。\n>\n> 原因：base 表是\"一个 case_id 一行\"的骨架,如果子表也是\"一个 case_id 一行\",可以直接 JOIN;如果子表是\"一个 case_id 多行\",JOIN 后会行数爆炸\n>\n> 查看：rows / unique_case_id；= 1.0\t一个 case_id 一行，不聚合✕ ,直接 JOIN；> 1.0\t一个 case_id 多行，✅ 必须聚合\n\n\n（1）只做类型压缩(Int64 → UInt32/UInt8,Date 字符串 → Date)，辅助字段不加！！\n\n（2）后期建朴素特征表用：base_clean_train.parquet ⭐️\n\n<img src=\"attachment:b4e00214-2502-40a0-bbf8-3ab32674f4c8.png\" width=\"30%\" />","metadata":{},"attachments":{"b4e00214-2502-40a0-bbf8-3ab32674f4c8.png":{"image/png":"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"}}},{"cell_type":"code","source":"# 数据集路径\nfrom pathlib import Path\n\n# 查看 /kaggle/input 下有哪些数据集\ninput_root = Path(\"/kaggle/input\")\nprint(\"=== /kaggle/input 下的数据集 ===\")\nfor p in sorted(input_root.iterdir()):\n    print(f\"  {p.name}\")\n\nprint(\"\\n=== 递归查找所有 base 文件 ===\")\nfor p in input_root.rglob(\"*base*.parquet\"):\n    print(f\"  {p}\")\n\nprint(\"\\n=== 递归查找所有 base csv 文件(备用)===\")\nfor p in input_root.rglob(\"*base*.csv\"):\n    print(f\"  {p}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.148957Z","iopub.execute_input":"2026-08-23T10:25:35.149399Z","iopub.status.idle":"2026-08-23T10:25:35.198488Z","shell.execute_reply.started":"2026-08-23T10:25:35.149365Z","shell.execute_reply":"2026-08-23T10:25:35.197330Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 环境准备\n# ============================================================\nimport polars as pl\nimport numpy as np\nfrom pathlib import Path\nimport gc\nimport shutil\nfrom datetime import datetime\n\n# 竞赛数据根目录(Kaggle 环境,注意多一层 competitions)\nRAW_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability\")\n\n# 输出目录(Kaggle 只能写 /kaggle/working)\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# base 表源路径(train + test,用 parquet 格式)\nTRAIN_BASE_PATH = RAW_ROOT / \"parquet_files\" / \"train\" / \"train_base.parquet\"\nTEST_BASE_PATH  = RAW_ROOT / \"parquet_files\" / \"test\"  / \"test_base.parquet\"\n\n# 输出路径\nOUT_BASE_PATH = OUT_ROOT / \"base_clean.parquet\"\nLOG_PATH      = OUT_ROOT / \"base_clean_log.txt\"\n\n# 校验路径存在\nprint(f\"训练 base 存在: {TRAIN_BASE_PATH.exists()}\")\nprint(f\"测试 base 存在: {TEST_BASE_PATH.exists()}\")\nprint(f\"输出目录: {OUT_ROOT}\")\n\nassert TRAIN_BASE_PATH.exists(), f\"❌ 找不到 {TRAIN_BASE_PATH}\"\nassert TEST_BASE_PATH.exists(),  f\"❌ 找不到 {TEST_BASE_PATH}\"\nprint(\"\\n✅ 路径全部正确,可以继续 Cell 2\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.200404Z","iopub.execute_input":"2026-08-23T10:25:35.200858Z","iopub.status.idle":"2026-08-23T10:25:35.213542Z","shell.execute_reply.started":"2026-08-23T10:25:35.200796Z","shell.execute_reply":"2026-08-23T10:25:35.212620Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: Step 0 - 备份原始 base 文件到 working 目录\n# 说明: Kaggle 的 /kaggle/input 是只读的,不会被误删,\n#      但为了流程完整性,复制一份到 working/backup 便于追溯\n# ============================================================\nBACKUP_DIR = OUT_ROOT / \"backup_raw\"\nBACKUP_DIR.mkdir(parents=True, exist_ok=True)\n\nshutil.copy(TRAIN_BASE_PATH, BACKUP_DIR / \"train_base_raw.parquet\")\nshutil.copy(TEST_BASE_PATH,  BACKUP_DIR / \"test_base_raw.parquet\")\n\nprint(\"✅ 备份完成:\")\nfor f in BACKUP_DIR.iterdir():\n    print(f\"  {f.name}  ({f.stat().st_size / 1024**2:.2f} MB)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.215017Z","iopub.execute_input":"2026-08-23T10:25:35.215405Z","iopub.status.idle":"2026-08-23T10:25:35.245505Z","shell.execute_reply.started":"2026-08-23T10:25:35.215358Z","shell.execute_reply":"2026-08-23T10:25:35.244432Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: Step 1 - 加载 base 表并校验 case_id 唯一性\n# ============================================================\ntrain_base = pl.read_parquet(TRAIN_BASE_PATH)\ntest_base  = pl.read_parquet(TEST_BASE_PATH)\n\nprint(f\"train_base shape: {train_base.shape}\")\nprint(f\"test_base  shape: {test_base.shape}\")\nprint(f\"\\ntrain columns: {train_base.columns}\")\nprint(f\"test  columns: {test_base.columns}\")\n\n# --- 唯一性校验 ---\ntrain_n_rows   = train_base.height\ntrain_n_unique = train_base[\"case_id\"].n_unique()\ntest_n_rows    = test_base.height\ntest_n_unique  = test_base[\"case_id\"].n_unique()\n\nprint(\"\\n\" + \"=\"*50)\nprint(\"case_id 唯一性校验\")\nprint(\"=\"*50)\nprint(f\"train: {train_n_rows} 行 / {train_n_unique} 唯一 case_id \"\n      f\"→ {'✅ 完全唯一' if train_n_rows == train_n_unique else f'⚠️ 有 {train_n_rows - train_n_unique} 个重复'}\")\nprint(f\"test : {test_n_rows} 行 / {test_n_unique} 唯一 case_id \"\n      f\"→ {'✅ 完全唯一' if test_n_rows == test_n_unique else f'⚠️ 有 {test_n_rows - test_n_unique} 个重复'}\")\n\n# 存到全局变量供后续 log 使用\nuniqueness_check = {\n    \"train_n_rows\": train_n_rows,\n    \"train_n_unique\": train_n_unique,\n    \"test_n_rows\": test_n_rows,\n    \"test_n_unique\": test_n_unique,\n}","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.247244Z","iopub.execute_input":"2026-08-23T10:25:35.248131Z","iopub.status.idle":"2026-08-23T10:25:35.432351Z","shell.execute_reply.started":"2026-08-23T10:25:35.248097Z","shell.execute_reply":"2026-08-23T10:25:35.431063Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 5: Step 2 - 类型转换(降内存 + 规范化日期)\n# ============================================================\ndef cast_base_types(df: pl.DataFrame, has_target: bool) -> pl.DataFrame:\n    \"\"\"\n    统一的类型转换函数,train/test 都用它\n    - case_id  : Int64 → UInt32(max 约 270 万,足够)\n    - MONTH    : Int64 → UInt32(202001 这种,UInt16 存不下)\n    - WEEK_NUM : Int64 → UInt8(max=91,足够)\n    - target   : Int64 → UInt8(只有 0/1)【test 没这列,跳过】\n    - date_decision : String → Date\n    \"\"\"\n    exprs = [\n        pl.col(\"case_id\").cast(pl.UInt32),\n        pl.col(\"MONTH\").cast(pl.UInt32),\n        pl.col(\"WEEK_NUM\").cast(pl.UInt8),\n        pl.col(\"date_decision\").str.to_date(\"%Y-%m-%d\"),\n    ]\n    if has_target:\n        exprs.append(pl.col(\"target\").cast(pl.UInt8))\n    return df.with_columns(exprs)\n\n# 转换前的内存\nmem_before_train = train_base.estimated_size(\"mb\")\nmem_before_test  = test_base.estimated_size(\"mb\")\n\ntrain_base = cast_base_types(train_base, has_target=True)\ntest_base  = cast_base_types(test_base,  has_target=False)\n\n# 转换后的内存\nmem_after_train = train_base.estimated_size(\"mb\")\nmem_after_test  = test_base.estimated_size(\"mb\")\n\nprint(\"类型转换完成\")\nprint(f\"train 内存: {mem_before_train:.2f} MB → {mem_after_train:.2f} MB \"\n      f\"(降低 {(1 - mem_after_train/mem_before_train)*100:.1f}%)\")\nprint(f\"test  内存: {mem_before_test:.2f} MB → {mem_after_test:.2f} MB \"\n      f\"(降低 {(1 - mem_after_test/mem_before_test)*100:.1f}%)\")\n\nprint(\"\\n转换后的 dtypes:\")\nprint(train_base.schema)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.433740Z","iopub.execute_input":"2026-08-23T10:25:35.434067Z","iopub.status.idle":"2026-08-23T10:25:35.492539Z","shell.execute_reply.started":"2026-08-23T10:25:35.434038Z","shell.execute_reply":"2026-08-23T10:25:35.491369Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6(替换版): 不合并 train/test,各自保留\n# 说明: 采用 Pattern B(分离处理),不加 data_source 字段\n# ============================================================\n\n# train: 保持原样(含 target),不加 data_source\ntrain_base_clean = train_base.select([\n    \"case_id\", \"date_decision\", \"MONTH\", \"WEEK_NUM\", \"target\"\n])\n\n# test: 不含 target,不加 data_source\ntest_base_clean = test_base.select([\n    \"case_id\", \"date_decision\", \"MONTH\", \"WEEK_NUM\"\n])\n\nprint(f\"train_base_clean shape: {train_base_clean.shape}\")\nprint(f\"test_base_clean  shape: {test_base_clean.shape}\")\nprint(f\"\\ntrain schema: {train_base_clean.schema}\")\nprint(f\"test  schema: {test_base_clean.schema}\")\nprint(f\"\\ntrain 内存: {train_base_clean.estimated_size('mb'):.2f} MB\")\nprint(f\"test  内存: {test_base_clean.estimated_size('mb'):.2f} MB\")\nprint(f\"\\n前 3 行 train:\\n{train_base_clean.head(3)}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.494920Z","iopub.execute_input":"2026-08-23T10:25:35.495265Z","iopub.status.idle":"2026-08-23T10:25:35.504731Z","shell.execute_reply.started":"2026-08-23T10:25:35.495234Z","shell.execute_reply":"2026-08-23T10:25:35.503712Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 7(替换版): 分别导出 train / test 两个 parquet 文件\n# 说明: write_parquet 会自动覆盖同名文件,无需手动删除\n# ============================================================\n\n# 定义新的输出路径(替换掉原来的合并路径)\nOUT_TRAIN_PATH = OUT_ROOT / \"base_clean_train.parquet\"\nOUT_TEST_PATH  = OUT_ROOT / \"base_clean_test.parquet\"\n\n# 写入(自动覆盖)\ntrain_base_clean.write_parquet(OUT_TRAIN_PATH, compression=\"zstd\", compression_level=3)\ntest_base_clean.write_parquet(OUT_TEST_PATH,  compression=\"zstd\", compression_level=3)\n\ntrain_size_mb = OUT_TRAIN_PATH.stat().st_size / 1024**2\ntest_size_mb  = OUT_TEST_PATH.stat().st_size  / 1024**2\n\nprint(f\"✅ 已导出 train: {OUT_TRAIN_PATH}  ({train_size_mb:.2f} MB)\")\nprint(f\"✅ 已导出 test : {OUT_TEST_PATH}   ({test_size_mb:.2f} MB)\")\n\n# --- 清理旧的合并文件(如果存在)---\nold_merged_path = OUT_ROOT / \"base_clean.parquet\"\nif old_merged_path.exists():\n    old_merged_path.unlink()\n    print(f\"\\n🗑️ 已删除旧的合并文件: {old_merged_path.name}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.506097Z","iopub.execute_input":"2026-08-23T10:25:35.506437Z","iopub.status.idle":"2026-08-23T10:25:35.638653Z","shell.execute_reply.started":"2026-08-23T10:25:35.506397Z","shell.execute_reply":"2026-08-23T10:25:35.637575Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 8(替换版): 更新处理日志\n# ============================================================\nlog_content = f\"\"\"\n============================================================\nbase 表清洗日志(Pattern B: train/test 分离处理)\n生成时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n\n【1】数据来源\n  train: {TRAIN_BASE_PATH}\n  test : {TEST_BASE_PATH}\n\n【2】唯一性校验\n  train 原始行数    : {uniqueness_check['train_n_rows']:,}\n  train 唯一 case_id: {uniqueness_check['train_n_unique']:,}\n  train 是否有重复  : {'否' if uniqueness_check['train_n_rows'] == uniqueness_check['train_n_unique'] else '是'}\n  \n  test  原始行数    : {uniqueness_check['test_n_rows']:,}\n  test  唯一 case_id: {uniqueness_check['test_n_unique']:,}\n  test  是否有重复  : {'否' if uniqueness_check['test_n_rows'] == uniqueness_check['test_n_unique'] else '是'}\n\n【3】类型转换\n  case_id       : Int64  → UInt32\n  MONTH         : Int64  → UInt32\n  WEEK_NUM      : Int64  → UInt8\n  target        : Int64  → UInt8 (仅 train)\n  date_decision : String → Date\n\n【4】处理策略:Pattern B(train / test 分离)\n  ✗ 不合并 train / test\n  ✗ 不添加 data_source 字段\n  ✓ 后续特征工程时,同一个函数分别应用到 train 和 test\n  \n  理由:\n    1) 真实 test 由 Kaggle 后台隐藏,本地 test 只有 10 行,合并意义有限\n    2) 竞赛属于未来外推场景,分离处理可天然避免数据泄漏\n    3) 流程更清晰,更贴近生产环境思维\n\n  ※ fold_id 和 is_covid_period 推迟到朴素特征表建好后添加\n\n【5】输出\n  train 文件: {OUT_TRAIN_PATH.name}  ({train_size_mb:.2f} MB, {train_base_clean.height:,} 行)\n  test  文件: {OUT_TEST_PATH.name}   ({test_size_mb:.2f} MB, {test_base_clean.height:,} 行)\n  压缩方式: zstd (level 3)\n\n【6】最终 schema\n  train: {train_base_clean.schema}\n  test : {test_base_clean.schema}\n============================================================\n\"\"\"\n\nwith open(LOG_PATH, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\n\nprint(log_content)\nprint(f\"\\n✅ 日志已覆盖保存到: {LOG_PATH}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.639910Z","iopub.execute_input":"2026-08-23T10:25:35.640213Z","iopub.status.idle":"2026-08-23T10:25:35.649949Z","shell.execute_reply.started":"2026-08-23T10:25:35.640185Z","shell.execute_reply":"2026-08-23T10:25:35.649022Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 9(替换版): 释放内存\n# ============================================================\ndel train_base, test_base, train_base_clean, test_base_clean\ndel uniqueness_check\ngc.collect()\n\nprint(\"✅ 内存已释放\")\n\n# 查看 working 目录当前状态\nprint(f\"\\n当前 {OUT_ROOT} 内容:\")\nfor f in sorted(OUT_ROOT.rglob(\"*\")):\n    if f.is_file():\n        size = f.stat().st_size / 1024**2\n        print(f\"  {f.relative_to(OUT_ROOT)}  ({size:.2f} MB)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.651241Z","iopub.execute_input":"2026-08-23T10:25:35.651648Z","iopub.status.idle":"2026-08-23T10:25:35.773237Z","shell.execute_reply.started":"2026-08-23T10:25:35.651582Z","shell.execute_reply":"2026-08-23T10:25:35.771958Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# static_0 表\n\n**（1）只做清洗：**\n\n每个 case_id 严格只有 1 行——所以 static_0 完全不需要聚合,只需要\"清洗\"\n\n**（2）清洗核心目标**\n\n只删 4 个字段、日期转 Date、M 后缀转 Categorical。分块给出,最后释放内存 \n\n删掉无信息量的列(全空、常数)：\n\n\n> deferredmnthsnum_166L（允许延期还款的月数）：它是常数(nunique=1)=可能Home Credit 的所有贷款产品都是同一个宽限期(比如都是 0,不给宽限)；这个字段设计好了但从未启用,永远填默认值不管哪种,每个客户的这个值都相同 → 无区分度 → 数学上就是 0 信息量）数据表里经常有一些字段是业务上定义了但实际未启用的,通过 nunique 分析可以识别出来。风控建模第一步就是清理这类无信号字段,避免引入噪声。\n> \n> lastrepayingdate_696D (缺失率: 99.84%)\n>\n> payvacationpostpone_4187118D (缺失率: 99.38%)：这两个字段是日期（Date）类型，分别代表“最后还款日期”和“最后一次还款假期的分期日期”()。在缺失率如此极端的情况下，它们已经失去了作为时间分布特征的价值，可以直接无脑删除。对于机器学习模型来说，缺失率超过 99% 的字段意味着全表只有不到 1% 的样本有观测值，这不仅无法提供泛化规律，反而极易导致模型对少数样本过拟合。\n>\n> typesuite_864L (缺失率: 73.46%)：该字段被算法标记为 constant_candidate（常数候选）；；数据显示其 approx_n_unique 仅为 1，在非空的抽样观测值中全部都是 'AL'()。这意味着该列除了缺失值（NaN）之外，只存在一种分类状态。如果业务上不需要将 NaN 作为特定状态（比如：是否有随行人员），它在数学上就是零方差的废字段，可以直接砍掉。\n\n\n类型转换(String 日期 → Date,降内存)\n\n保留所有可能有用的字段,不做业务派生(留到朴素表后)\n\n**（3）删除清单：**\n![屏幕截图 2026-08-21 184058.png](attachment:5e68151b-2259-46c1-b47d-2afe847cb397.png)\n\n# ----------------------------------------------------------------------\n\n**（4）后期优化：**\n\n输出文件 165 MB，清洗动作没错,但压缩效果没达到预期。原因是 Kaggle 上 Polars 的某些默认行为导致 zstd 压缩不理想。朴素表建成后再统一优化内存也来得及！","metadata":{},"attachments":{"5e68151b-2259-46c1-b47d-2afe847cb397.png":{"image/png":"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"}}},{"cell_type":"code","source":"# ============================================================\n# 通用工具:test schema 强制对齐 train (整个 notebook 定义一次)\n# ============================================================\ndef align_test_to_train(test_df: pl.DataFrame, train_parquet_path) -> pl.DataFrame:\n    \"\"\"\n    按 train 的完整 schema 对齐 test:\n    - train 有 test 没有的列:数值型填 0,其他填 null\n    - test 有 train 没有的列:直接丢弃\n    - 最终按 train 的列名、顺序、dtype 强制对齐\n    - 对齐函数只动 test,不碰 train\n    \"\"\"\n    train_schema = pl.read_parquet(train_parquet_path, n_rows=0).schema\n    train_cols = list(train_schema.keys())\n    \n    NUMERIC_DTYPES = {\n        pl.Int8, pl.Int16, pl.Int32, pl.Int64,\n        pl.UInt8, pl.UInt16, pl.UInt32, pl.UInt64,\n        pl.Float32, pl.Float64,\n    }\n    \n    # 1. 补齐 test 缺失的列\n    for col in train_cols:\n        if col not in test_df.columns:\n            dtype = train_schema[col]\n            fill_val = 0 if dtype in NUMERIC_DTYPES else None\n            test_df = test_df.with_columns(pl.lit(fill_val).cast(dtype).alias(col))\n    \n    # 2. 丢弃 test 多出来的列\n    extra_cols = [c for c in test_df.columns if c not in train_schema]\n    if extra_cols:\n        test_df = test_df.drop(extra_cols)\n    \n    # 3. 按 train 的列顺序 + dtype 强制对齐\n    test_df = test_df.select([pl.col(c).cast(train_schema[c]) for c in train_cols])\n    \n    print(f\"✅ Schema 对齐完成: 补 {len(train_cols) - test_df.width + len([c for c in train_cols if c not in test_df.columns])} 列 / 丢 {len(extra_cols)} 列 / 最终 {test_df.width} 列\")\n    return test_df\n\nprint(\"✅ align_test_to_train() 已定义,后续所有 test 处理都可以复用\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.774747Z","iopub.execute_input":"2026-08-23T10:25:35.775101Z","iopub.status.idle":"2026-08-23T10:25:35.789067Z","shell.execute_reply.started":"2026-08-23T10:25:35.775067Z","shell.execute_reply":"2026-08-23T10:25:35.788017Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 环境准备(如果 base 表清洗后 kernel 没重启,可跳过 import)\n# ============================================================\nimport polars as pl\nfrom pathlib import Path\nimport gc\nfrom datetime import datetime\n\n# 竞赛数据根目录\nRAW_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability\")\n\n# 输出目录\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# static_0 源路径(2 个分片)\nSTATIC0_SHARDS = [\n    RAW_ROOT / \"parquet_files\" / \"train\" / \"train_static_0_0.parquet\",\n    RAW_ROOT / \"parquet_files\" / \"train\" / \"train_static_0_1.parquet\",\n]\n\n# 输出路径\nOUT_STATIC0_PATH = OUT_ROOT / \"static_0_clean_train.parquet\"\nLOG_PATH = OUT_ROOT / \"static_0_clean_log.txt\"\n\n# 校验路径\nfor p in STATIC0_SHARDS:\n    assert p.exists(), f\"❌ 找不到 {p}\"\n    print(f\"✅ 存在: {p.name}  ({p.stat().st_size / 1024**2:.2f} MB)\")\n\nprint(f\"\\n输出目录: {OUT_ROOT}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.791793Z","iopub.execute_input":"2026-08-23T10:25:35.792157Z","iopub.status.idle":"2026-08-23T10:25:35.819936Z","shell.execute_reply.started":"2026-08-23T10:25:35.792124Z","shell.execute_reply":"2026-08-23T10:25:35.818497Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: 加载两个分片并纵向拼接\n# 说明: 探查显示 schema_consistent=True,可以直接 concat\n# ============================================================\nshard_0 = pl.read_parquet(STATIC0_SHARDS[0])\nshard_1 = pl.read_parquet(STATIC0_SHARDS[1])\n\nprint(f\"shard_0 shape: {shard_0.shape}\")\nprint(f\"shard_1 shape: {shard_1.shape}\")\n\n# 校验 schema 一致(否则 concat 会报错)\nassert shard_0.schema == shard_1.schema, \"❌ 两个分片 schema 不一致!\"\n\n# 纵向拼接\nstatic_0 = pl.concat([shard_0, shard_1], how=\"vertical\")\nprint(f\"\\n拼接后 shape: {static_0.shape}\")\nprint(f\"总内存: {static_0.estimated_size('mb'):.2f} MB\")\n\n# 校验 case_id 唯一性(depth=0 应该一对一)\nn_rows = static_0.height\nn_unique = static_0[\"case_id\"].n_unique()\nprint(f\"\\ncase_id 唯一性检查: {n_rows} 行 / {n_unique} 唯一 \"\n      f\"→ {'✅ 一对一' if n_rows == n_unique else f'⚠️ 有 {n_rows - n_unique} 重复'}\")\n\n# 释放分片\ndel shard_0, shard_1\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:35.821545Z","iopub.execute_input":"2026-08-23T10:25:35.822139Z","iopub.status.idle":"2026-08-23T10:25:37.167437Z","shell.execute_reply.started":"2026-08-23T10:25:35.822082Z","shell.execute_reply":"2026-08-23T10:25:37.166427Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: 删除 4 个明确无信号的字段\n# ============================================================\nCOLS_TO_DROP = [\n    \"deferredmnthsnum_166L\",         # 常数(nunique=1),业务上定义但从未启用\n    \"lastrepayingdate_696D\",         # 缺失率 99.84%,时间分布特征失效\n    \"payvacationpostpone_4187118D\",  # 缺失率 99.38%,时间分布特征失效\n    \"typesuite_864L\",                # 缺失率 73.46%,非空全为 'AL',零方差\n]\n\n# 记录删除前列数\nn_cols_before = static_0.width\n\n# 校验待删列都存在(防止拼错)\nmissing = [c for c in COLS_TO_DROP if c not in static_0.columns]\nassert len(missing) == 0, f\"❌ 以下列不在表中: {missing}\"\n\n# 执行删除\nstatic_0 = static_0.drop(COLS_TO_DROP)\n\nprint(f\"删除前列数: {n_cols_before}\")\nprint(f\"删除后列数: {static_0.width}\")\nprint(f\"实际删除: {n_cols_before - static_0.width} 列\")\nprint(f\"\\n被删除的字段:\")\nfor c in COLS_TO_DROP:\n    print(f\"  - {c}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:37.168554Z","iopub.execute_input":"2026-08-23T10:25:37.168884Z","iopub.status.idle":"2026-08-23T10:25:37.182187Z","shell.execute_reply.started":"2026-08-23T10:25:37.168856Z","shell.execute_reply":"2026-08-23T10:25:37.181102Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 4: String 日期字段 → Date 类型\n# 说明: \n#   - D 后缀 = 日期字段(Home Credit 命名规范)\n#   - 但源数据里存的是 String,需要 str.to_date 转换\n#   - 缺失保持 null,不填充\n# ============================================================\n\n# 找出所有 D 后缀且当前是 String 类型的列\ndate_cols_to_convert = [\n    c for c in static_0.columns\n    if c.endswith(\"D\") and static_0[c].dtype == pl.String\n]\n\nprint(f\"识别到 {len(date_cols_to_convert)} 个 String 类型的日期字段需要转换:\")\nfor c in date_cols_to_convert:\n    null_rate = static_0[c].null_count() / static_0.height\n    print(f\"  - {c}  (缺失率: {null_rate:.2%})\")\n\n# 批量转换\nstatic_0 = static_0.with_columns([\n    pl.col(c).str.to_date(\"%Y-%m-%d\", strict=False).alias(c)\n    for c in date_cols_to_convert\n])\n\nprint(f\"\\n✅ 已转换 {len(date_cols_to_convert)} 个字段为 Date 类型\")\n\n# 验证:抽查一个字段的类型\nif date_cols_to_convert:\n    sample_col = date_cols_to_convert[0]\n    print(f\"抽查 {sample_col} 类型: {static_0[sample_col].dtype}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:37.183829Z","iopub.execute_input":"2026-08-23T10:25:37.184421Z","iopub.status.idle":"2026-08-23T10:25:37.645293Z","shell.execute_reply.started":"2026-08-23T10:25:37.184373Z","shell.execute_reply":"2026-08-23T10:25:37.644332Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 5: M 后缀字段(加密掩码类别)→ Categorical 类型\n# 说明:\n#   - M 后缀 = Masking 加密类别(如 'a55475b1')\n#   - String → Categorical 可节省 60-80% 内存\n#   - 加速后续 groupby 和 join\n# ============================================================\n\n# 找出所有 M 后缀且当前是 String 类型的列\nmask_cols_to_convert = [\n    c for c in static_0.columns\n    if c.endswith(\"M\") and static_0[c].dtype == pl.String\n]\n\nprint(f\"识别到 {len(mask_cols_to_convert)} 个 M 后缀字段需要转 Categorical:\")\nfor c in mask_cols_to_convert:\n    n_uniq = static_0[c].n_unique()\n    print(f\"  - {c}  (unique 值数: {n_uniq})\")\n\n# 记录转换前内存\nmem_before = static_0.estimated_size(\"mb\")\n\n# 批量转换\nstatic_0 = static_0.with_columns([\n    pl.col(c).cast(pl.Categorical).alias(c)\n    for c in mask_cols_to_convert\n])\n\nmem_after = static_0.estimated_size(\"mb\")\nprint(f\"\\n✅ 已转换 {len(mask_cols_to_convert)} 个 M 字段为 Categorical\")\nprint(f\"内存: {mem_before:.2f} MB → {mem_after:.2f} MB \"\n      f\"(降低 {(1 - mem_after/mem_before)*100:.1f}%)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:37.646447Z","iopub.execute_input":"2026-08-23T10:25:37.646779Z","iopub.status.idle":"2026-08-23T10:25:38.053900Z","shell.execute_reply.started":"2026-08-23T10:25:37.646749Z","shell.execute_reply":"2026-08-23T10:25:38.053034Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6: case_id 转 UInt32(和 base 表保持一致,方便后续 JOIN)\n# ============================================================\n\n# case_id 转 UInt32(与 base 表对齐,防止 JOIN 时因类型不同报错)\nstatic_0 = static_0.with_columns(\n    pl.col(\"case_id\").cast(pl.UInt32)\n)\n\nprint(f\"case_id dtype: {static_0['case_id'].dtype}\")\n\n# --- 最终检查 ---\nprint(\"\\n\" + \"=\"*60)\nprint(\"最终表结构概览\")\nprint(\"=\"*60)\nprint(f\"总行数: {static_0.height:,}\")\nprint(f\"总列数: {static_0.width}\")\nprint(f\"总内存: {static_0.estimated_size('mb'):.2f} MB\")\n\n# 按数据类型统计列数\ndtype_counts = {}\nfor col, dtype in static_0.schema.items():\n    dtype_str = str(dtype)\n    dtype_counts[dtype_str] = dtype_counts.get(dtype_str, 0) + 1\n\nprint(f\"\\n按类型统计列数:\")\nfor dt, cnt in sorted(dtype_counts.items(), key=lambda x: -x[1]):\n    print(f\"  {dt}: {cnt} 列\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:38.054783Z","iopub.execute_input":"2026-08-23T10:25:38.055159Z","iopub.status.idle":"2026-08-23T10:25:38.073430Z","shell.execute_reply.started":"2026-08-23T10:25:38.055130Z","shell.execute_reply":"2026-08-23T10:25:38.071710Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 7: 导出 static_0_clean_train.parquet\n# ============================================================\nstatic_0.write_parquet(\n    OUT_STATIC0_PATH,\n    compression=\"zstd\",\n    compression_level=3,\n)\n\nfile_size_mb = OUT_STATIC0_PATH.stat().st_size / 1024**2\nprint(f\"✅ 已导出: {OUT_STATIC0_PATH}\")\nprint(f\"   文件大小: {file_size_mb:.2f} MB\")\nprint(f\"   总行数  : {static_0.height:,}\")\nprint(f\"   总列数  : {static_0.width}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:38.075006Z","iopub.execute_input":"2026-08-23T10:25:38.075403Z","iopub.status.idle":"2026-08-23T10:25:46.917979Z","shell.execute_reply.started":"2026-08-23T10:25:38.075369Z","shell.execute_reply":"2026-08-23T10:25:46.916790Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# 【补做】static_0 的 test 清洗 (复用 train 的所有清洗规则)\n# ============================================================\nTEST_STATIC0_SHARDS = [\n    RAW_ROOT / \"parquet_files\" / \"test\" / \"test_static_0_0.parquet\",\n    RAW_ROOT / \"parquet_files\" / \"test\" / \"test_static_0_1.parquet\",\n    RAW_ROOT / \"parquet_files\" / \"test\" / \"test_static_0_2.parquet\",\n]\nOUT_STATIC0_TEST_PATH = OUT_ROOT / \"static_0_clean_test.parquet\"\n\n# --- Step 1: 加载并纵向拼接(用 vertical_relaxed 兼容全 null 列的 dtype 差异)---\ntest_static_0 = pl.concat(\n    [pl.read_parquet(p) for p in TEST_STATIC0_SHARDS],\n    how=\"vertical_relaxed\"\n)\nprint(f\"test_static_0 拼接后 shape: {test_static_0.shape}\")\n\n# --- Step 2: 复用 train 的删除清单(容错:只删存在的列)---\ndrop_now = [c for c in COLS_TO_DROP if c in test_static_0.columns]\ntest_static_0 = test_static_0.drop(drop_now)\n\n# --- Step 3: D 后缀日期字段转换 ---\nfor c in test_static_0.columns:\n    if c.endswith(\"D\") and test_static_0[c].dtype == pl.String:\n        test_static_0 = test_static_0.with_columns(\n            pl.col(c).str.to_date(\"%Y-%m-%d\", strict=False)\n        )\n\n# --- Step 4: M 后缀转 Categorical ---\nfor c in test_static_0.columns:\n    if c.endswith(\"M\") and test_static_0[c].dtype == pl.String:\n        test_static_0 = test_static_0.with_columns(pl.col(c).cast(pl.Categorical))\n\n# --- Step 5: case_id 转 UInt32 ---\ntest_static_0 = test_static_0.with_columns(pl.col(\"case_id\").cast(pl.UInt32))\n\n# --- Step 6: 强制对齐 train schema(防御性,static_0 理论上列数一致)---\ntest_static_0 = align_test_to_train(test_static_0, OUT_STATIC0_PATH)\n\n# --- Step 7: 导出 ---\ntest_static_0.write_parquet(\n    OUT_STATIC0_TEST_PATH,\n    compression=\"zstd\",\n    compression_level=3,\n)\nprint(f\"✅ 已导出: {OUT_STATIC0_TEST_PATH}\")\nprint(f\"   shape: {test_static_0.shape}\")\n\ndel test_static_0\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:46.919357Z","iopub.execute_input":"2026-08-23T10:25:46.919776Z","iopub.status.idle":"2026-08-23T10:25:47.102469Z","shell.execute_reply.started":"2026-08-23T10:25:46.919735Z","shell.execute_reply":"2026-08-23T10:25:47.101587Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 8: 生成 static_0 清洗日志 (含 train + test 双份记录)\n# ============================================================\n\n# --- 从 parquet 反推所有变量,不依赖内存中已被 del 的中间变量 ---\nOUT_STATIC0_TEST_PATH = OUT_ROOT / \"static_0_clean_test.parquet\"\n\ntrain_schema_check = pl.read_parquet(OUT_STATIC0_PATH, n_rows=0).schema\ntest_schema_check  = pl.read_parquet(OUT_STATIC0_TEST_PATH, n_rows=0).schema\n\ntrain_n_rows = pl.read_parquet(OUT_STATIC0_PATH, columns=[\"case_id\"]).height\ntest_n_rows  = pl.read_parquet(OUT_STATIC0_TEST_PATH, columns=[\"case_id\"]).height\ntrain_n_cols = len(train_schema_check)\ntest_n_cols  = len(test_schema_check)\n\ntrain_file_size_mb = OUT_STATIC0_PATH.stat().st_size / 1024**2\ntest_file_size_mb  = OUT_STATIC0_TEST_PATH.stat().st_size / 1024**2\n\nschema_align_status = \"✅ 完全一致\" if train_schema_check == test_schema_check else \"❌ 不一致\"\n\ndate_cols_to_convert = [c for c, dt in train_schema_check.items() if c.endswith(\"D\") and dt == pl.Date]\nmask_cols_to_convert = [c for c, dt in train_schema_check.items() if c.endswith(\"M\") and dt == pl.Categorical]\n\nlog_content = f\"\"\"\n============================================================\nstatic_0 表清洗日志\n生成时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n【1】数据来源\n  ── Train ──\n  分片 1: train_static_0_0.parquet\n  分片 2: train_static_0_1.parquet\n  拼接方式: 纵向 concat(schema 一致性校验通过)\n  ── Test ──\n  分片 1: test_static_0_0.parquet\n  分片 2: test_static_0_1.parquet\n  分片 3: test_static_0_2.parquet\n  拼接方式: vertical_relaxed (兼容全空列的 dtype 差异)\n\n【2】case_id 唯一性\n  Train 总行数    : {train_n_rows:,}\n  Test  总行数    : {test_n_rows:,}\n  结论: depth=0 表,一对一,无需聚合,仅做字段清洗\n\n【3】删除的无信号字段(共 4 个,train/test 应用相同规则)\n  - deferredmnthsnum_166L\n    理由: 常数字段(nunique=1),业务上定义但从未启用,零信息量\n  \n  - lastrepayingdate_696D\n    理由: 缺失率 99.84%,时间分布特征失效\n  \n  - payvacationpostpone_4187118D\n    理由: 缺失率 99.38%,时间分布特征失效\n  \n  - typesuite_864L\n    理由: 缺失率 73.46% + 非空全为 'AL',零方差常数候选\n\n【4】类型转换 (train/test 应用相同规则)\n  D 后缀日期字段 ({len(date_cols_to_convert)} 个): String → Date\n    转换字段: {date_cols_to_convert}\n  \n  M 后缀掩码字段 ({len(mask_cols_to_convert)} 个): String → Categorical\n    转换字段: {mask_cols_to_convert}\n  \n  case_id: Int64 → UInt32 (与 base 表对齐)\n\n【5】保留策略\n  ✗ 未删除 near_constant_in_sample 字段\n    理由: 近常数 ≠ 常数,可能是\"是否触发\"型强特征,保留\n  \n  ✗ 未做数值列的类型下探(Float64 保持不变)\n    理由: 第一版优先保证正确性,后续可再优化内存\n\n【6】输出\n  ── Train ──\n  文件路径: {OUT_STATIC0_PATH}\n  文件大小: {train_file_size_mb:.2f} MB\n  总行数  : {train_n_rows:,}\n  总列数  : {train_n_cols}\n  \n  ── Test ──\n  文件路径: {OUT_STATIC0_TEST_PATH}\n  文件大小: {test_file_size_mb:.4f} MB\n  总行数  : {test_n_rows:,}\n  总列数  : {test_n_cols}\n  \n  压缩方式: zstd (level 3)\n  Schema 对齐: {schema_align_status}\n\n【7】最终 schema(仅列出前 20 列)\n\"\"\"\nfor col, dtype in list(train_schema_check.items())[:20]:\n    log_content += f\"  {col}: {dtype}\\n\"\nlog_content += f\"  ...(共 {train_n_cols} 列)\\n\"\nlog_content += \"=\"*60 + \"\\n\"\n\nstatic_0_log_path = OUT_ROOT / \"static_0_clean_log.txt\"\nwith open(static_0_log_path, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\nprint(f\"\\n✅ 日志已保存到: {static_0_log_path}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:47.103692Z","iopub.execute_input":"2026-08-23T10:25:47.104028Z","iopub.status.idle":"2026-08-23T10:25:47.136477Z","shell.execute_reply.started":"2026-08-23T10:25:47.103989Z","shell.execute_reply":"2026-08-23T10:25:47.135401Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 9: 释放内存,为处理下一张表腾空间\n# ============================================================\ndel static_0\ndel date_cols_to_convert, mask_cols_to_convert\ngc.collect()\n\nprint(\"✅ 内存已释放\")\n\n# 查看 working/clean 目录当前状态\nprint(f\"\\n当前 {OUT_ROOT} 内容:\")\nfor f in sorted(OUT_ROOT.rglob(\"*\")):\n    if f.is_file():\n        size = f.stat().st_size / 1024**2\n        print(f\"  {f.relative_to(OUT_ROOT)}  ({size:.2f} MB)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:47.137951Z","iopub.execute_input":"2026-08-23T10:25:47.138277Z","iopub.status.idle":"2026-08-23T10:25:47.226583Z","shell.execute_reply.started":"2026-08-23T10:25:47.138246Z","shell.execute_reply":"2026-08-23T10:25:47.225498Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 10: 快速验证清洗结果\n# ============================================================\nimport polars as pl\nfrom pathlib import Path\n\nOUT_STATIC0_PATH = Path(\"/kaggle/working/clean/static_0_clean_train.parquet\")\ndf = pl.read_parquet(OUT_STATIC0_PATH)\n\nprint(\"=\"*60)\nprint(\"【验证 1】列数是否符合预期\")\nprint(\"=\"*60)\nprint(f\"预期: 168 - 4 = 164 列\")\nprint(f\"实际: {df.width} 列\")\nprint(f\"结果: {'✅' if df.width == 164 else '❌'}\")\n\nprint(\"\\n\" + \"=\"*60)\nprint(\"【验证 2】4 个删除字段是否真的不存在\")\nprint(\"=\"*60)\nshould_be_dropped = [\n    \"deferredmnthsnum_166L\",\n    \"lastrepayingdate_696D\",\n    \"payvacationpostpone_4187118D\",\n    \"typesuite_864L\",\n]\nfor c in should_be_dropped:\n    exists = c in df.columns\n    print(f\"  {c}: {'❌ 还在' if exists else '✅ 已删'}\")\n\nprint(\"\\n\" + \"=\"*60)\nprint(\"【验证 3】剩下的 9 个 String 列是什么\")\nprint(\"=\"*60)\nstring_cols = [c for c, dt in df.schema.items() if dt == pl.String]\nprint(f\"共 {len(string_cols)} 个 String 列:\")\nfor c in string_cols:\n    n_uniq = df[c].n_unique()\n    null_rate = df[c].null_count() / df.height\n    sample = df[c].drop_nulls().head(3).to_list()\n    print(f\"  {c}\")\n    print(f\"    unique={n_uniq}, null_rate={null_rate:.2%}, 样本值={sample}\")\n\nprint(\"\\n\" + \"=\"*60)\nprint(\"【验证 4】Date 列样例(确认时间范围合理)\")\nprint(\"=\"*60)\ndate_cols = [c for c, dt in df.schema.items() if dt == pl.Date]\nfor c in date_cols[:5]:  # 只看前 5 个\n    min_d = df[c].min()\n    max_d = df[c].max()\n    print(f\"  {c}: {min_d} ~ {max_d}\")\n\ndel df\nimport gc\ngc.collect()\nprint(\"\\n✅ 验证完成\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:47.228043Z","iopub.execute_input":"2026-08-23T10:25:47.228438Z","iopub.status.idle":"2026-08-23T10:25:48.716906Z","shell.execute_reply.started":"2026-08-23T10:25:47.228397Z","shell.execute_reply":"2026-08-23T10:25:48.715881Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# static_cb_0 表\n\n（1）只做清洗：\n\n每个 case_id 只有 1 行,所以完全不需要聚合,只需要清洗字段；这和 base 表一样是 depth=0,一 case 一行的结构。工作重心是丢弃无信息量的列 + 类型转换。\n\n（2）字段初筛判定：\n\n✅ 独特值数 = 1(纯常数)\t删除\n\n⚠️ 独特值数 ≥ 2,不管缺失率多高\t保留\n\n理由：即使 97.6% 缺失,只要剩下 2.4% 里有 47 个不同值,这 2.4% 的样本对模型来说就是强稀疏信号。LightGBM/XGBoost 天然能处理稀疏,不需要我们提前筛。\n\n（3）只删 3列(纯常数)，这 3 个是无争议的垃圾——独特值只有 1 个意味着这一列所有非空值全部相同,即使全保留也是常数,模型学不到任何东西,连稀疏信号都算不上\n\nformonth_118L\t97.6%\t1\t纯常数,零信息\n\nforquarter_462L\t97.6%\t1\t纯常数,零信息\n\nforquarter_601L(应为 forweek_601L)\t97.6%\t1\t纯常数,零信息\n\n# -----------------------------------------------------\n\n（4）日期字段类型转换(String → Date)\n\n从字段名后缀 D 和探查表识别出 9 个日期列:\n\n列名                 原类型    目标类型\n\nassignmentdate_238D\tString\tDate\n\nassignmentdate_4527235D\tString\tDate\n\nassignmentdate_4955616D\tString\tDate\n\nbirthdate_574D\tString\tDate\n\ndateofbirth_337D\tString\tDate\n\ndateofbirth_342D\tString\tDate\n\nresponsedate_1012D\tString\tDate\n\nresponsedate_4527233D\tString\tDate\n\nresponsedate_4917613D\tString\tDate\n\n# -------------------------------------------------------\n\n（5）case_id 类型压缩\n\n列名\t原类型\t目标类型\t依据\n\ncase_id\tInt64\tUInt32\tmax = 2,703,454 < 2³² ≈ 42 亿\n\n# ------------------------------------------------------\n\n（6）其他字段保持原类型不变\n\n所有 Float64 数值列 → 暂不压缩(不同列范围差异大,统一处理反而危险)\n\n所有 String 类别列(以 M 结尾)→ 暂不转 Categorical(Kaggle 上 Polars 的 Categorical 跨表拼接容易出问题,朴素表建好后再统一编码)\n\n所有 Boolean 列 → 保持\n# ------------------------------------------------------\n\n（7）\n\n\n\n\n\n","metadata":{}},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 环境准备\n# ============================================================\nimport polars as pl\nfrom pathlib import Path\nimport gc\nfrom datetime import datetime\n\nRAW_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability\")\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# 源路径\nTRAIN_SRC = RAW_ROOT / \"parquet_files\" / \"train\" / \"train_static_cb_0.parquet\"\nTEST_SRC  = RAW_ROOT / \"parquet_files\" / \"test\"  / \"test_static_cb_0.parquet\"\n\n# 输出路径\nOUT_TRAIN = OUT_ROOT / \"static_cb_0_clean_train.parquet\"\nOUT_TEST  = OUT_ROOT / \"static_cb_0_clean_test.parquet\"\nLOG_PATH  = OUT_ROOT / \"static_cb_0_clean_log.txt\"\n\nprint(f\"train 源存在: {TRAIN_SRC.exists()}\")\nprint(f\"test  源存在: {TEST_SRC.exists()}\")\n\nassert TRAIN_SRC.exists() and TEST_SRC.exists(), \"❌ 源文件路径不对\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:48.717998Z","iopub.execute_input":"2026-08-23T10:25:48.718259Z","iopub.status.idle":"2026-08-23T10:25:48.729286Z","shell.execute_reply.started":"2026-08-23T10:25:48.718235Z","shell.execute_reply":"2026-08-23T10:25:48.728269Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: 加载 train / test,基本检查\n# ============================================================\ntrain_cb = pl.read_parquet(TRAIN_SRC)\ntest_cb  = pl.read_parquet(TEST_SRC)\n\nprint(f\"train shape: {train_cb.shape}\")\nprint(f\"test  shape: {test_cb.shape}\")\nprint(f\"\\ntrain 列数: {len(train_cb.columns)}\")\nprint(f\"test  列数: {len(test_cb.columns)}\")\n\n# 校验列名一致(除了 target,static_cb_0 里没 target)\nassert set(train_cb.columns) == set(test_cb.columns), \"⚠️ train/test 列不一致!\"\nprint(\"\\n✅ train / test 列名完全一致\")\n\n# 校验 case_id 唯一性\ntrain_unique = train_cb[\"case_id\"].n_unique()\ntest_unique  = test_cb[\"case_id\"].n_unique()\nprint(f\"\\ntrain: {train_cb.height} 行 / {train_unique} 唯一 case_id \"\n      f\"→ {'✅ 唯一' if train_cb.height == train_unique else '⚠️ 有重复'}\")\nprint(f\"test : {test_cb.height} 行 / {test_unique} 唯一 case_id \"\n      f\"→ {'✅ 唯一' if test_cb.height == test_unique else '⚠️ 有重复'}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:48.730476Z","iopub.execute_input":"2026-08-23T10:25:48.730840Z","iopub.status.idle":"2026-08-23T10:25:48.918692Z","shell.execute_reply.started":"2026-08-23T10:25:48.730796Z","shell.execute_reply":"2026-08-23T10:25:48.917728Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: 定义要删除的列 + 要转 Date 的列\n# ============================================================\n\n# 【动作 1】要删除的纯常数列\nDROP_COLS = [\n    \"formonth_118L\",\n    \"forquarter_462L\",\n    \"forweek_601L\",\n]\n\n# 【动作 2】要从 String 转 Date 的列(所有 D 后缀日期列)\nDATE_COLS = [\n    \"assignmentdate_238D\",\n    \"assignmentdate_4527235D\",\n    \"assignmentdate_4955616D\",\n    \"birthdate_574D\",\n    \"dateofbirth_337D\",\n    \"dateofbirth_342D\",\n    \"responsedate_1012D\",\n    \"responsedate_4527233D\",\n    \"responsedate_4917613D\",\n]\n\n# 校验:确保这些列都存在于 train 里\nmissing_drop = [c for c in DROP_COLS if c not in train_cb.columns]\nmissing_date = [c for c in DATE_COLS if c not in train_cb.columns]\nassert not missing_drop, f\"❌ 待删除列不存在: {missing_drop}\"\nassert not missing_date, f\"❌ 待转日期列不存在: {missing_date}\"\n\nprint(f\"✅ 待删除列 {len(DROP_COLS)} 个,全部存在\")\nprint(f\"✅ 待转日期列 {len(DATE_COLS)} 个,全部存在\")\nprint(f\"\\n清洗后预计列数: {len(train_cb.columns) - len(DROP_COLS)} = \"\n      f\"{len(train_cb.columns)} - {len(DROP_COLS)}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:48.920114Z","iopub.execute_input":"2026-08-23T10:25:48.920406Z","iopub.status.idle":"2026-08-23T10:25:48.928456Z","shell.execute_reply.started":"2026-08-23T10:25:48.920378Z","shell.execute_reply":"2026-08-23T10:25:48.927284Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 4(修正版): 定义健壮的清洗函数\n# 修复: 兼容 test 表里因全 NaN 被推断为 Float64 的日期列\n# ============================================================\ndef clean_static_cb(df: pl.DataFrame) -> pl.DataFrame:\n    \"\"\"\n    对 static_cb_0 表做最小侵入式清洗:\n      1. 删除 3 个纯常数列\n      2. case_id: Int64 → UInt32\n      3. 9 个日期列: String → Date(如果原本是 String)\n                   Float64(全 NaN) → Date null(如果原本是 Float64)\n    \"\"\"\n    # Step 1: 删除常数列\n    df = df.drop(DROP_COLS)\n    \n    # Step 2: case_id 类型压缩\n    df = df.with_columns(pl.col(\"case_id\").cast(pl.UInt32))\n    \n    # Step 3: 日期列处理(按当前类型分派)\n    date_exprs = []\n    for c in DATE_COLS:\n        current_dtype = df.schema[c]\n        if current_dtype == pl.String:\n            # 正常情况: String → Date\n            date_exprs.append(pl.col(c).str.to_date(\"%Y-%m-%d\", strict=False))\n        else:\n            # 异常情况: 全 NaN 被推断为 Float64 → 直接转 Date null\n            date_exprs.append(pl.lit(None).cast(pl.Date).alias(c))\n    \n    df = df.with_columns(date_exprs)\n    return df\n\n# 记录清洗前状态\ntrain_shape_before = train_cb.shape\ntest_shape_before  = test_cb.shape\ntrain_mem_before   = train_cb.estimated_size(\"mb\")\ntest_mem_before    = test_cb.estimated_size(\"mb\")\n\nprint(\"清洗前:\")\nprint(f\"  train: {train_shape_before}, {train_mem_before:.2f} MB\")\nprint(f\"  test : {test_shape_before}, {test_mem_before:.2f} MB\")\n\n# 打印 test 里各日期列的当前类型,帮你确认到底是哪些列被推断错了\nprint(\"\\ntest 表中日期列的当前 dtype:\")\nfor c in DATE_COLS:\n    print(f\"  {c}: {test_cb.schema[c]}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:48.929840Z","iopub.execute_input":"2026-08-23T10:25:48.930188Z","iopub.status.idle":"2026-08-23T10:25:48.995788Z","shell.execute_reply.started":"2026-08-23T10:25:48.930158Z","shell.execute_reply":"2026-08-23T10:25:48.994609Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 5: 应用清洗函数\n# ============================================================\ntrain_cb_clean = clean_static_cb(train_cb)\ntest_cb_clean  = clean_static_cb(test_cb)\n\ntrain_shape_after = train_cb_clean.shape\ntest_shape_after  = test_cb_clean.shape\ntrain_mem_after   = train_cb_clean.estimated_size(\"mb\")\ntest_mem_after    = test_cb_clean.estimated_size(\"mb\")\n\nprint(\"清洗后:\")\nprint(f\"  train: {train_shape_after}, {train_mem_after:.2f} MB \"\n      f\"(降低 {(1 - train_mem_after/train_mem_before)*100:.1f}%)\")\nprint(f\"  test : {test_shape_after}, {test_mem_after:.2f} MB \"\n      f\"(降低 {(1 - test_mem_after/test_mem_before)*100:.1f}%)\")\n\n# 校验:被删的列真的没了、日期列真的是 Date 了\nfor c in DROP_COLS:\n    assert c not in train_cb_clean.columns, f\"❌ {c} 未删除\"\nfor c in DATE_COLS:\n    assert train_cb_clean.schema[c] == pl.Date, f\"❌ {c} 未转 Date\"\nassert train_cb_clean.schema[\"case_id\"] == pl.UInt32, \"❌ case_id 未转 UInt32\"\n\nprint(\"\\n✅ 所有清洗校验通过\")\n\n# 抽 3 行看看日期转换效果\nprint(\"\\n转换后的日期列样例(前 3 行,选几个日期列):\")\nsample_date_cols = [\"case_id\", \"dateofbirth_337D\", \"responsedate_1012D\", \"birthdate_574D\"]\nprint(train_cb_clean.select(sample_date_cols).head(3))","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:49.000652Z","iopub.execute_input":"2026-08-23T10:25:49.001026Z","iopub.status.idle":"2026-08-23T10:25:49.188763Z","shell.execute_reply.started":"2026-08-23T10:25:49.000991Z","shell.execute_reply":"2026-08-23T10:25:49.187514Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6: 导出为 parquet(自动覆盖同名文件)\n# ============================================================\ntrain_cb_clean.write_parquet(OUT_TRAIN, compression=\"zstd\", compression_level=3)\ntest_cb_clean.write_parquet(OUT_TEST,  compression=\"zstd\", compression_level=3)\n\ntrain_size_mb = OUT_TRAIN.stat().st_size / 1024**2\ntest_size_mb  = OUT_TEST.stat().st_size  / 1024**2\n\nprint(f\"✅ 已导出 train: {OUT_TRAIN.name}  ({train_size_mb:.2f} MB, {train_cb_clean.height:,} 行)\")\nprint(f\"✅ 已导出 test : {OUT_TEST.name}   ({test_size_mb:.2f} MB, {test_cb_clean.height:,} 行)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:49.190174Z","iopub.execute_input":"2026-08-23T10:25:49.190728Z","iopub.status.idle":"2026-08-23T10:25:50.663583Z","shell.execute_reply.started":"2026-08-23T10:25:49.190645Z","shell.execute_reply":"2026-08-23T10:25:50.662589Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 7: 写处理日志\n# ============================================================\nlog_content = f\"\"\"\n============================================================\nstatic_cb_0 表清洗日志\n生成时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n\n【1】数据来源\n  train: {TRAIN_SRC}\n  test : {TEST_SRC}\n\n【2】表性质\n  ✓ depth=0 外部征信局静态数据\n  ✓ one_row_case_rate = 1.0,每个 case_id 只有一行\n  ✓ 无需聚合,可直接 LEFT JOIN 到 base 表\n\n【3】唯一性校验\n  train: {train_cb.height:,} 行 / {train_unique:,} 唯一 case_id\n  test : {test_cb.height:,} 行 / {test_unique:,} 唯一 case_id\n\n【4】清洗原则\n  ✗ 独特值 = 1 (纯常数) → 删除\n  ✓ 独特值 ≥ 2 (即使 97% 缺失) → 全部保留\n    理由: 稀疏但独特的值对 LightGBM 是强稀疏信号\n  ✗ 不做聚合(每个 case_id 只有一行)\n  ✗ 不做特征派生(距 date_decision 天数、is_null 标记等留到朴素表阶段)\n\n【5】动作 1:删除 3 个纯常数列\n  - formonth_118L    (独特值=1)\n  - forquarter_462L  (独特值=1)\n  - forweek_601L     (独特值=1)\n\n【6】动作 2:类型转换\n  case_id : Int64 → UInt32\n  以下 9 个日期列: String → Date\n{chr(10).join(f'    - {c}' for c in DATE_COLS)}\n\n【7】其他字段\n  - Float64 数值列: 保持原类型(不同列范围差异大,统一压缩风险高)\n  - String 类别列 (M 后缀 hash): 保持 String,待朴素表建好后统一编码\n  - Boolean 列: 保持原样\n\n【8】清洗前后对比\n  train shape: {train_shape_before} → {train_shape_after}\n  train 内存: {train_mem_before:.2f} MB → {train_mem_after:.2f} MB\n  test  shape: {test_shape_before} → {test_shape_after}\n  test  内存: {test_mem_before:.2f} MB → {test_mem_after:.2f} MB\n\n【9】输出\n  train 文件: {OUT_TRAIN.name}  ({train_size_mb:.2f} MB)\n  test  文件: {OUT_TEST.name}   ({test_size_mb:.2f} MB)\n  压缩方式: zstd (level 3)\n\n【10】最终 schema(train,共 {len(train_cb_clean.columns)} 列)\n{train_cb_clean.schema}\n============================================================\n\"\"\"\n\nwith open(LOG_PATH, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\n\nprint(log_content)\nprint(f\"\\n✅ 日志已保存到: {LOG_PATH}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:50.665886Z","iopub.execute_input":"2026-08-23T10:25:50.666834Z","iopub.status.idle":"2026-08-23T10:25:50.676675Z","shell.execute_reply.started":"2026-08-23T10:25:50.666785Z","shell.execute_reply":"2026-08-23T10:25:50.675687Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 8: 释放内存\n# ============================================================\ndel train_cb, test_cb, train_cb_clean, test_cb_clean\ngc.collect()\n\nprint(\"✅ 内存已释放\")\n\n# 查看 working/clean 目录当前状态\nprint(f\"\\n当前 {OUT_ROOT} 内容:\")\nfor f in sorted(OUT_ROOT.rglob(\"*\")):\n    if f.is_file():\n        size = f.stat().st_size / 1024**2\n        print(f\"  {f.relative_to(OUT_ROOT)}  ({size:.2f} MB)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:50.678201Z","iopub.execute_input":"2026-08-23T10:25:50.678644Z","iopub.status.idle":"2026-08-23T10:25:51.065266Z","shell.execute_reply.started":"2026-08-23T10:25:50.678579Z","shell.execute_reply":"2026-08-23T10:25:51.064271Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# applprev_1 表\n\n（1）表结构详述：\n\n总行数\t6,525,979 行(2 个分片)\t大表,需要 Polars lazy\n\n总列数\t41 列\t数值 + 类别 + 日期混合\n\n**主键\tcase_id + num_group1\t每个客户有多条历史申请**\n\ncase_id 唯一数\t1,212,667\t只覆盖 79% 的 base 客户(base 有 152 万)\n\n每个 case_id 平均记录数\t5.34 条,p90=12,p99=20,max=20\t大部分人历史申请不多,但有长尾\n\n只有 1 条记录的比例\t18.8%\t说明大多数人有多条记录,聚合必须做\n\n业务含义\t客户在本机构的历史申请记录\t每一行是\"过去某次申请\"\n# ------------------------------------------------------\n\n（2）键理解:num_group1 的语义：\n\nnum_group1=0, 1, 2, 3, ... 是同一个 case_id 下的多次历史申请\n\n但顺序不一定是时间顺序(例:case_id=6 的 num_group1=0 是 2018 年,num_group1=2 是 2014 年)，真正的时间顺序要靠 creationdate_885D 字段\n\n这个理解很重要,直接决定后面\"最近一次申请\"怎么取！\n\n# ------------------------------------------------------\n\n（3）真正达到\"绝对垃圾\"标准的只有 1 个：\n\nprofession_152M\t缺失率 0%,但 11,558 个唯一值,全是掩码字符串,a55475b1 一个值就占了绝大多数,基本是无效数据。可以只保留 nunique 一个聚合特征作为纪念,原字段丢弃。\n# ------------------------------------------------------\n\n（4）Step 1:类型转换\n\n所有 _D 结尾字段 String → Date\n\n所有 Int64 键 → UInt32\n\n所有 Float64 数值字段保留(不能压 Float32,会丢精度)\n\nBoolean → UInt8(0/1)\n# ------------------------------------------------------\n\n（5）模型修正：applprev_1\n\npivoted = cat_df.pivot(\n    values=\"cnt\",\n    index=\"case_id\",\n    on=col,\n)\n\npivot 的意思是:把 col 这一列里\"实际出现过的每一个值\"变成一个新列。\n\n项目最终会 JOIN 9 张表,总列数上千。如果不在单表阶段解决,等到跨表合并阶段:\n\n出错时你不知道是哪张表引入的\n不知道应该填 0 还是 null(某些列填 0 是错的,比如金额均值)\n不知道dtype 应该是什么\n\n这就是软件工程里的**\"错误就近处理\"原则**——问题在哪个模块产生,就在哪个模块内解决。\n\n![屏幕截图 2026-08-23 171708.png](attachment:083e3503-c135-4495-89cd-a50762f7dd05.png)\n# ------------------------------------------------------\n\n（6）static_0 的问题是完全不同的性质：\n\nstatic_0 不是 pivot 问题,是流程遗漏——你的原代码从 Cell 14 到 Cell 22 只处理了 train,压根没写 test 那部分。\n\n# ------------------------------------------------------\n\n在机器学习工程里有一个铁律:\n\n训练集定义 schema,测试集必须无条件遵从。\n\n这不是我发明的,是所有生产 ML 系统的默认约定。原因很简单——模型是在 train 上学的,test 只能按 train 的样子来,反过来不行。\n\n所以对齐函数的逻辑就三条:\n\nTrain 有的列,Test 必须有 → 缺就补\nTest 有但 Train 没有的列 → 无用信息,直接扔\nTest 列的顺序和 dtype → 完全对齐 train","metadata":{},"attachments":{"083e3503-c135-4495-89cd-a50762f7dd05.png":{"image/png":"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"}}},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 路径与环境准备 (Kaggle 环境)\n# ============================================================\nimport polars as pl\nimport gc\nimport time\nimport json\nimport os\nfrom pathlib import Path\nfrom datetime import datetime\n\n# 1. 定义数据源根目录 (Kaggle Dataset 路径)\nDATA_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability/parquet_files\")\nTRAIN_DIR = DATA_ROOT / \"train\"\nTEST_DIR = DATA_ROOT / \"test\"\n\n# 2. 获取 applprev_1 的所有分片路径 (自动匹配 train_applprev_1_0.parquet, 1_1.parquet 等)\nTRAIN_APPLPREV_PATHS = sorted(list(TRAIN_DIR.glob(\"train_applprev_1_*.parquet\")))\nTEST_APPLPREV_PATHS = sorted(list(TEST_DIR.glob(\"test_applprev_1_*.parquet\")))\n\n# 3. 定义输出目录 (放到 clean 文件夹下)\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# 4. 定义聚合后文件的保存路径 (对齐命名规范)\nOUT_TRAIN_PATH = OUT_ROOT / \"applprev_1_clean_train.parquet\"\nOUT_TEST_PATH = OUT_ROOT / \"applprev_1_clean_test.parquet\"\n\n# 4. 定义聚合后文件的保存路径\nOUT_TRAIN_PATH = OUT_ROOT / \"train_applprev_1_agg.parquet\"\nOUT_TEST_PATH = OUT_ROOT / \"test_applprev_1_agg.parquet\"\n\n# 验证路径是否正确提取\nprint(f\"✅ 输出目录已准备: {OUT_ROOT}\")\nprint(f\"🔍 找到 train applprev_1 分片数: {len(TRAIN_APPLPREV_PATHS)}\")\nfor p in TRAIN_APPLPREV_PATHS:\n    print(f\"   - {p.name}\")\n    \nprint(f\"🔍 找到 test applprev_1 分片数 : {len(TEST_APPLPREV_PATHS)}\")\nfor p in TEST_APPLPREV_PATHS:\n    print(f\"   - {p.name}\")\n\n# 如果找不到文件，触发断言报错，避免往后跑空代码\nassert len(TRAIN_APPLPREV_PATHS) > 0, \"❌ 没找到 train_applprev_1 文件，请检查路径！\"\nassert len(TEST_APPLPREV_PATHS) > 0, \"❌ 没找到 test_applprev_1 文件，请检查路径！\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:51.067299Z","iopub.execute_input":"2026-08-23T10:25:51.068018Z","iopub.status.idle":"2026-08-23T10:25:51.086287Z","shell.execute_reply.started":"2026-08-23T10:25:51.067977Z","shell.execute_reply":"2026-08-23T10:25:51.085120Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: 字段分类字典 —— 所有聚合规则的配置中心\n# ============================================================\n\n# --- 主键字段 ---\nKEY_COLS = [\"case_id\", \"num_group1\"]\n\n# --- 要丢弃的字段(只有 profession_152M，均为掩码无效数据)---\nDROP_COLS = [\"profession_152M\"]\n\n# --- 数值字段(P/A/L 后缀,除 num_group1 外)---\n# 这些字段做全套聚合:max/min/mean/sum/std\nNUMERIC_COLS = [\n    \"actualdpd_943P\",              # 实际逾期天数\n    \"annuity_853A\",                # 月供\n    \"byoccupationinc_3656910L\",    # 历史申请中收入(缺失76%,但保留)\n    \"childnum_21L\",                # 子女数\n    \"credacc_actualbalance_314A\",  # 缺失95%,保留\n    \"credacc_credlmt_575A\",        # 信用额度\n    \"credacc_maxhisbal_375A\",      # 缺失95%,保留\n    \"credacc_minhisbal_90A\",       # 缺失95%,保留\n    \"credacc_transactions_402L\",   # 缺失95%,保留\n    \"credamount_590A\",             # 贷款金额\n    \"currdebt_94A\",                # 当前债务\n    \"downpmt_134A\",                # 首付\n    \"mainoccupationinc_437A\",      # 主要收入\n    \"maxdpdtolerance_577P\",        # 最大DPD容忍\n    \"outstandingdebt_522A\",        # 未偿债务\n    \"pmtnum_8L\",                   # 付款次数\n    \"revolvingaccount_394A\",       # 循环账户(缺失95%,保留)\n    \"tenor_203L\",                  # 分期期数\n]\n\n# --- DPD 类字段(需要额外的风险计数特征)---\nDPD_COLS = [\"actualdpd_943P\", \"maxdpdtolerance_577P\"]\n\n# --- 布尔字段 ---\nBOOL_COLS = [\n    \"isbidproduct_390L\",           # 是否交叉销售\n    \"isdebitcard_527L\",            # 是否借记卡(缺失93%)\n]\n\n# --- 低基数类别字段(做 count + pct 编码)---\nLOW_CARD_CAT_COLS = [\n    \"credtype_587L\",               # 3 个值:CAL/COL/REL\n    \"status_219L\",                 # 11 个值:D/T/A/K/N...\n    \"inittransactioncode_279L\",    # 3 个值:CASH/POS/NDF\n    \"familystate_726L\",            # 5 个值\n    \"credacc_status_367L\",         # 6 个值,缺失95%\n]\n\n# --- 中基数类别字段(只做 nunique)---\nMID_CARD_CAT_COLS = [\n    \"cancelreason_3545846M\",       # 79 个值\n    \"postype_4733339M\",            # 9 个值\n    \"rejectreason_755M\",           # 19 个值\n    \"rejectreasonclient_4145042M\", # 14 个值\n    \"education_1138M\",             # 6 个值\n]\n\n# --- 高基数类别字段(只做 nunique)---\nHIGH_CARD_CAT_COLS = [\n    \"district_544M\",               # 1130 个值\n]\n\n# --- 日期字段 ---\nDATE_COLS = [\n    \"creationdate_885D\",           # 关键!排序基准\n    \"approvaldate_319D\",           # 缺失46%\n    \"dateactivated_425D\",          # 缺失48%\n    \"dtlastpmt_581D\",              # 缺失73%\n    \"dtlastpmtallstes_3545839D\",   # 缺失62%\n    \"employedfrom_700D\",           # 缺失60%\n    \"firstnonzeroinstldate_307D\",  # 缺失10%\n]\n\n# --- \"最近一次申请\"快照要保留的字段(补偿聚合信息损失)---\nLAST_SNAPSHOT_COLS = [\n    \"credamount_590A\",\n    \"annuity_853A\",\n    \"credtype_587L\",\n    \"status_219L\",\n    \"actualdpd_943P\",\n    \"outstandingdebt_522A\",\n    \"tenor_203L\",\n    \"mainoccupationinc_437A\",\n]\n\nprint(f\"字段分类完成:\")\nprint(f\"  丢弃字段    : {len(DROP_COLS)}\")\nprint(f\"  数值字段    : {len(NUMERIC_COLS)}\")\nprint(f\"  DPD 字段    : {len(DPD_COLS)}\")\nprint(f\"  布尔字段    : {len(BOOL_COLS)}\")\nprint(f\"  低基数类别  : {len(LOW_CARD_CAT_COLS)}\")\nprint(f\"  中基数类别  : {len(MID_CARD_CAT_COLS)}\")\nprint(f\"  高基数类别  : {len(HIGH_CARD_CAT_COLS)}\")\nprint(f\"  日期字段    : {len(DATE_COLS)}\")\nprint(f\"  快照字段    : {len(LAST_SNAPSHOT_COLS)}\")\n\n# 校验:所有字段覆盖(除了目标列和其他无用主键)\nall_covered = set(KEY_COLS + NUMERIC_COLS + BOOL_COLS + LOW_CARD_CAT_COLS +\n                  MID_CARD_CAT_COLS + HIGH_CARD_CAT_COLS + DATE_COLS + DROP_COLS)\nprint(f\"\\n配置覆盖字段总数(应为41): {len(all_covered)}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:51.087751Z","iopub.execute_input":"2026-08-23T10:25:51.088071Z","iopub.status.idle":"2026-08-23T10:25:51.102162Z","shell.execute_reply.started":"2026-08-23T10:25:51.088025Z","shell.execute_reply":"2026-08-23T10:25:51.100872Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: 核心聚合函数 —— train / test 共用这一个函数 (仅针对 applprev_1)\n# ============================================================\n\ndef build_applprev_single_table_features(\n    applprev_paths: list,\n    verbose: bool = True,\n) -> pl.DataFrame:\n    \"\"\"\n    单表聚合清理 applprev_1 (不强依赖 base 表，只针对有历史申请的客户)\n    \"\"\"\n    \n# --- Step 1: 读取分片并合并 ---\n    if verbose:\n        print(f\"[Step 1] 读取 {len(applprev_paths)} 个分片...\")\n    \n    df_list = [pl.scan_parquet(p) for p in applprev_paths]\n    \n    # 🌟 关键修改：使用 vertical_relaxed 解决空分片导致的 Null 类型推断冲突\n    df = pl.concat(df_list, how=\"vertical_relaxed\")\n\n    \n    # --- Step 2: 丢弃绝对垃圾字段 ---\n    df = df.drop(DROP_COLS)\n    \n    # --- Step 3: 类型转换 ---\n    if verbose:\n        print(f\"[Step 2] 类型转换(日期 String→Date, 主键降级 UInt32)...\")\n    \n    df = df.with_columns([\n        pl.col(\"case_id\").cast(pl.UInt32),\n        pl.col(\"num_group1\").cast(pl.UInt16),\n    ])\n    for date_col in DATE_COLS:\n        df = df.with_columns(pl.col(date_col).str.to_date(\"%Y-%m-%d\", strict=False))\n    for bool_col in BOOL_COLS:\n        df = df.with_columns(pl.col(bool_col).cast(pl.UInt8))\n    \n    # --- Step 4: 日期衍生特征 (因为没有 base，计算内部的相对时长) ---\n    if verbose:\n        print(f\"[Step 3] 衍生特征 (approval_lag_days)...\")\n    \n    # 审批时长 = approval - creation\n    df = df.with_columns(\n        (pl.col(\"approvaldate_319D\") - pl.col(\"creationdate_885D\"))\n          .dt.total_days()\n          .cast(pl.Int32)\n          .alias(\"approval_lag_days\")\n    )\n    \n    # --- Step 5: 构造\"最近一次申请\"快照 ---\n    if verbose:\n        print(f\"[Step 4] 抽取每个 case_id 的最近一次申请快照...\")\n    \n    # 按 creationdate 降序,取第一条 (即最近的一次申请)\n    last_snapshot = (\n        df.sort([\"case_id\", \"creationdate_885D\"], descending=[False, True])\n          .group_by(\"case_id\")\n          .agg([pl.col(c).first().alias(f\"last_{c}\") for c in LAST_SNAPSHOT_COLS])\n    )\n    \n    # --- Step 6: 主聚合 ---\n    if verbose:\n        print(f\"[Step 5] 主聚合(按 case_id)...\")\n    \n    agg_exprs = []\n    \n    # 全局:总记录数\n    agg_exprs.append(pl.len().cast(pl.UInt16).alias(\"applprev_n_records\"))\n    \n    # 数值字段:max/min/mean/sum/std/count_non_null\n    for col in NUMERIC_COLS:\n        agg_exprs.extend([\n            pl.col(col).max().alias(f\"{col}_max\"),\n            pl.col(col).min().alias(f\"{col}_min\"),\n            pl.col(col).mean().alias(f\"{col}_mean\"),\n            pl.col(col).sum().alias(f\"{col}_sum\"),\n            pl.col(col).std().alias(f\"{col}_std\"),\n            pl.col(col).is_not_null().sum().cast(pl.UInt16).alias(f\"{col}_cnt\"),\n        ])\n    \n    # DPD 字段:额外的风险计数(>0 和 >30)\n    for col in DPD_COLS:\n        agg_exprs.extend([\n            pl.col(col).filter(pl.col(col) > 0).len().cast(pl.UInt16).alias(f\"{col}_gt0_cnt\"),\n            pl.col(col).filter(pl.col(col) > 30).len().cast(pl.UInt16).alias(f\"{col}_gt30_cnt\"),\n        ])\n    \n    # 布尔字段:sum + mean\n    for col in BOOL_COLS:\n        agg_exprs.extend([\n            pl.col(col).sum().cast(pl.UInt16).alias(f\"{col}_sum\"),\n            pl.col(col).mean().alias(f\"{col}_mean\"),\n        ])\n    \n    # 中/高基数类别:只做 nunique\n    for col in MID_CARD_CAT_COLS + HIGH_CARD_CAT_COLS:\n        agg_exprs.append(\n            pl.col(col).n_unique().cast(pl.UInt16).alias(f\"{col}_nunique\")\n        )\n    \n    # 日期字段 (单表没有基准点，提取极值即可，Polars 支持直接求日期的 max/min)\n    for date_col in DATE_COLS:\n        agg_exprs.extend([\n            pl.col(date_col).min().alias(f\"{date_col}_min\"),\n            pl.col(date_col).max().alias(f\"{date_col}_max\"),\n        ])\n    \n    # 衍生时长字段\n    agg_exprs.extend([\n        pl.col(\"approval_lag_days\").mean().alias(\"approval_lag_days_mean\"),\n        pl.col(\"approval_lag_days\").max().alias(\"approval_lag_days_max\"),\n    ])\n    \n    main_agg = df.group_by(\"case_id\").agg(agg_exprs)\n    \n    # --- Step 7: 低基数类别字段的 count/pct 编码 ---\n    if verbose:\n        print(f\"[Step 6] 低基数类别的 count/pct 编码配置...\")\n    \n    cat_aggs = []\n    for col in LOW_CARD_CAT_COLS:\n        pivot_expr = (\n            df.filter(pl.col(col).is_not_null())\n              .group_by([\"case_id\", col])\n              .agg(pl.len().alias(\"cnt\"))\n        )\n        cat_aggs.append((col, pivot_expr))\n    \n    # --- Step 8: 合并所有聚合结果 ---\n    if verbose:\n        print(f\"[Step 7] 执行计算(Collect)并合并主聚合与快照...\")\n    \n    result = main_agg.collect(streaming=True)\n    last_snap_df = last_snapshot.collect(streaming=True)\n    result = result.join(last_snap_df, on=\"case_id\", how=\"left\")\n    \n    # 处理低基数类别的透视\n    if verbose:\n        print(f\"[Step 8] 展开低基数类别透视并横向合并...\")\n    \n    for col, cat_lazy in cat_aggs:\n        cat_df = cat_lazy.collect(streaming=True)\n        if cat_df.height == 0:\n            continue\n        # 透视成宽表\n        pivoted = cat_df.pivot(\n            values=\"cnt\",\n            index=\"case_id\",\n            on=col,\n            aggregate_function=\"sum\",\n        ).fill_null(0)\n        \n        new_cols = [c for c in pivoted.columns if c != \"case_id\"]\n        pivoted = pivoted.rename({c: f\"{col}_{c}_cnt\" for c in new_cols})\n        \n        # 计算pct (比例)\n        cnt_cols = [f\"{col}_{c}_cnt\" for c in new_cols]\n        pivoted = pivoted.with_columns(\n            pl.sum_horizontal(cnt_cols).alias(f\"{col}_total\")\n        )\n        for c in new_cols:\n            pivoted = pivoted.with_columns(\n                (pl.col(f\"{col}_{c}_cnt\") / pl.col(f\"{col}_total\")).alias(f\"{col}_{c}_pct\")\n            )\n        pivoted = pivoted.drop(f\"{col}_total\")\n        \n        # Join 回主表\n        result = result.join(pivoted, on=\"case_id\", how=\"left\")\n    \n    # (不需要 Left Join 回 Base，单纯返回已清洗合并好的应用历史表)\n    if verbose:\n        print(f\"\\n✅ 单表清洗聚合完成: {result.shape}\")\n    \n    return result\n\nprint(\"✅ 单表清洗聚合函数已定义,可用于 Phase A/B/C\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:51.103749Z","iopub.execute_input":"2026-08-23T10:25:51.104186Z","iopub.status.idle":"2026-08-23T10:25:51.131972Z","shell.execute_reply.started":"2026-08-23T10:25:51.104138Z","shell.execute_reply":"2026-08-23T10:25:51.130934Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 4: Phase A - 小样本验证(提取前 100000 行跑通全流程)\n# ============================================================\nimport time\nimport gc\nimport polars as pl\n\nprint(\"生成单表小样本...\")\n# 直接从第一个分片截取 10w 行作为临时文件验证流程\nsample_applprev = pl.scan_parquet(TRAIN_APPLPREV_PATHS[0]).head(100000).collect()\n\nSAMPLE_APPLPREV_PATH = OUT_ROOT / \"_sample_applprev.parquet\"\nsample_applprev.write_parquet(SAMPLE_APPLPREV_PATH)\n\nt0 = time.time()\nsample_features = build_applprev_single_table_features(\n    applprev_paths=[SAMPLE_APPLPREV_PATH],\n    verbose=True,\n)\nelapsed = time.time() - t0\n\nprint(f\"\\n耗时: {elapsed:.1f} 秒\")\nprint(f\"输出 shape: {sample_features.shape}\")\nprint(f\"内存占用: {sample_features.estimated_size('mb'):.2f} MB\")\n\n# ⚠️ 注意事项提示\nif sample_features.width > 200:\n    print(f\"💡 提示: 输出的列数超过 200 ({sample_features.width}列) 是完全正常的！\")\n    print(\"   因为低基数类别(如 status_219L 等)经过透视(pivot)会展开成大量占比(pct)和计数(cnt)列。\")\n\nprint(f\"\\n前 3 行(部分列):\")\nprint(sample_features.head(3))\n\n# 清理临时文件\nSAMPLE_APPLPREV_PATH.unlink()\ndel sample_applprev, sample_features\ngc.collect()\n\nprint(\"\\n✅ Phase A 验证通过,可以进入 Phase B(全量 train)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:51.134502Z","iopub.execute_input":"2026-08-23T10:25:51.135333Z","iopub.status.idle":"2026-08-23T10:25:51.696122Z","shell.execute_reply.started":"2026-08-23T10:25:51.135279Z","shell.execute_reply":"2026-08-23T10:25:51.694797Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 5: Phase B - 全量 train 处理(约 6.5M 行)\n# ============================================================\n# 💡 注意: 全量运行时,如果内存不够,核心函数中的 main_agg.collect(streaming=True) \n#         会发挥作用。保持 streaming=True, Polars 会自动进行分批(out-of-core)处理。\n# ============================================================\nimport time\nimport gc\n\nprint(\"开始处理全量 train applprev_1...\")\nt0 = time.time()\n\ntrain_features = build_applprev_single_table_features(\n    applprev_paths=TRAIN_APPLPREV_PATHS,\n    verbose=True,\n)\n\nelapsed = time.time() - t0\nprint(f\"\\n耗时: {elapsed:.1f} 秒 ({elapsed/60:.1f} 分钟)\")\nprint(f\"输出 shape: {train_features.shape}\")\nprint(f\"内存占用: {train_features.estimated_size('mb'):.2f} MB\")\n\n# 导出纯净版单表特征 (无 base join)\ntrain_features.write_parquet(\n    OUT_TRAIN_PATH,\n    compression=\"zstd\",\n    compression_level=3,\n)\nfile_size_mb = OUT_TRAIN_PATH.stat().st_size / 1024**2\nprint(f\"\\n✅ 已导出: {OUT_TRAIN_PATH}\")\nprint(f\"   文件大小: {file_size_mb:.2f} MB\")\nprint(f\"   总行数  : {train_features.height:,}\")\nprint(f\"   总列数  : {train_features.width}\")\n\n# 释放全量数据内存\ndel train_features\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:25:51.697246Z","iopub.execute_input":"2026-08-23T10:25:51.697521Z","iopub.status.idle":"2026-08-23T10:26:23.198248Z","shell.execute_reply.started":"2026-08-23T10:25:51.697496Z","shell.execute_reply":"2026-08-23T10:26:23.197220Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6: Phase C - test 处理  \n# ============================================================\nimport time\nimport gc\nimport polars as pl\n\nprint(\"开始处理 test applprev_1...\")\nt0 = time.time()\n\ntest_features = build_applprev_single_table_features(\n    applprev_paths=TEST_APPLPREV_PATHS,\n    verbose=True,\n)\n\n\n# 🌟 新增这一行:强制对齐到 train schema\ntest_features = align_test_to_train(test_features, OUT_TRAIN_PATH)\n\n\nelapsed = time.time() - t0\nprint(f\"\\n耗时: {elapsed:.1f} 秒\")\nprint(f\"输出 shape: {test_features.shape}\")\n\n\ntest_features.write_parquet(\n    OUT_TEST_PATH,\n    compression=\"zstd\",\n    compression_level=3,\n)\nfile_size_mb = OUT_TEST_PATH.stat().st_size / 1024**2\nprint(f\"\\n✅ 已导出: {OUT_TEST_PATH}\")\nprint(f\"   文件大小: {file_size_mb:.4f} MB\")\nprint(f\"   总行数  : {test_features.height}\")\nprint(f\"   总列数  : {test_features.width}\")\n\n# 校验 train/test 列一致\ntrain_cols_list = pl.read_parquet(OUT_TRAIN_PATH, n_rows=1).columns\ntest_cols_list = test_features.columns\n\ntrain_cols = set(train_cols_list)\ntest_cols  = set(test_cols_list)\n\nmissing_in_test = train_cols - test_cols\nextra_in_test   = test_cols - train_cols\n\nprint(f\"\\ntrain 有但 test 没有的列: {len(missing_in_test)}\")\nprint(f\"test  有但 train 没有的列: {len(extra_in_test)}\")\n\n\n# 替换原来的 \"💡 提示\" 那几行 + if missing_in_test 判断\nif missing_in_test or extra_in_test:\n    print(\"\\n⚠️ Schema 仍不一致,请检查 align_test_to_train 是否生效\")\nelse:\n    print(\"\\n✅ Train/Test Schema 完全一致,可以进入下游流程\")\n\n\ndel test_features\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:23.199521Z","iopub.execute_input":"2026-08-23T10:26:23.200684Z","iopub.status.idle":"2026-08-23T10:26:23.506944Z","shell.execute_reply.started":"2026-08-23T10:26:23.200632Z","shell.execute_reply":"2026-08-23T10:26:23.506008Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 7: 写处理日志 (生成 .txt 并打印)\n# ============================================================\nfrom datetime import datetime\n\n# 构建纯文本日志内容\nlog_content = f\"\"\"============================================================\n数据清洗日志: applprev_1\n处理时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n\n[1. 字段丢弃]\n主动丢弃的无效/高缺失掩码字段 ({len(DROP_COLS)} 个):\n{', '.join(DROP_COLS)}\n\n[2. 训练集与测试集形态]\nTrain 特征列数 : {len(train_cols_list)} 列\nTest  特征列数 : {len(test_cols_list)} 列\n\n[3. Schema 差异 (通常由测试集极小导致的低基数透视缺失)]\nTrain 有但 Test 没有的列 ({len(missing_in_test)} 个):\n{', '.join(list(missing_in_test)) if missing_in_test else '无'}\n\nTest 有但 Train 没有的列 ({len(extra_in_test)} 个):\n{', '.join(list(extra_in_test)) if extra_in_test else '无'}\n\n============================================================\n\"\"\"\n\n# 打印日志到 Cell 输出\nprint(log_content)\n\n# 保存为 .txt 文件 (对齐命名规范)\nlog_path = OUT_ROOT / \"applprev_1_clean_log.txt\"\nwith open(log_path, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\n\nprint(f\"✅ 处理日志已保存至: {log_path}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:23.508273Z","iopub.execute_input":"2026-08-23T10:26:23.508803Z","iopub.status.idle":"2026-08-23T10:26:23.518064Z","shell.execute_reply.started":"2026-08-23T10:26:23.508698Z","shell.execute_reply":"2026-08-23T10:26:23.517019Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# applprev_2 表\n\n（1）历史申请的二级附属信息表，如联系方式、卡片状态等）\n\n（2）非常特殊：是典型的 Depth=2 表（通过 case_id + num_group1 + num_group2 定位），拥有高达 1400 万行的数据量，但只有 3 个业务字段，且全是类别型（String）\n\n（3）没有需要丢弃的字段：\n\ncredacc_cards_status_52L 虽然缺失率高达 97.5%，但在风控领域，这种“特定附属产品（信用卡）的状态”字段，缺失本身就代表了“没有办理该业务”的强信息\n\n而且它在非空时有 6 个有效独立值，并非像 profession_152M 那样只有一个无意义的掩码值。因此，我们目前全部保留，将其视作低基数类别进行全量展开（Count / Pct 编码）\n","metadata":{}},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 路径与环境准备 (applprev_2)\n# ============================================================\nimport polars as pl\nimport gc\nimport time\nfrom pathlib import Path\nfrom datetime import datetime\n\n# 1. 定义数据源根目录\nDATA_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability/parquet_files\")\nTRAIN_DIR = DATA_ROOT / \"train\"\nTEST_DIR = DATA_ROOT / \"test\"\n\n# 2. 获取 applprev_2 分片路径 (即使只有 1 个文件，也用 list 保持通用性)\nTRAIN_APPLPREV_2_PATHS = sorted(list(TRAIN_DIR.glob(\"train_applprev_2*.parquet\")))\nTEST_APPLPREV_2_PATHS = sorted(list(TEST_DIR.glob(\"test_applprev_2*.parquet\")))\n\n# 3. 定义 clean 输出目录\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# 4. 定义聚合后文件的保存路径\nOUT_TRAIN_PATH = OUT_ROOT / \"applprev_2_clean_train.parquet\"\nOUT_TEST_PATH = OUT_ROOT / \"applprev_2_clean_test.parquet\"\n\nassert len(TRAIN_APPLPREV_2_PATHS) > 0, \"❌ 没找到 train_applprev_2 文件！\"\nassert len(TEST_APPLPREV_2_PATHS) > 0, \"❌ 没找到 test_applprev_2 文件！\"\nprint(f\"✅ 环境准备完毕，准备处理 {len(TRAIN_APPLPREV_2_PATHS)} 个 Train 分片。\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:23.519381Z","iopub.execute_input":"2026-08-23T10:26:23.519756Z","iopub.status.idle":"2026-08-23T10:26:23.549054Z","shell.execute_reply.started":"2026-08-23T10:26:23.519711Z","shell.execute_reply":"2026-08-23T10:26:23.547030Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: 字段分类配置 (applprev_2 极简版)\n# ============================================================\n\n# --- 主键 ---\nKEY_COLS = [\"case_id\", \"num_group1\", \"num_group2\"]\n\n# --- 丢弃字段 (目前初筛认为无绝对垃圾) ---\nDROP_COLS = []\n\n# --- 低基数类别字段 (展开做 Count & Pct 编码) ---\nLOW_CARD_CAT_COLS = [\n    \"cacccardblochreas_147M\",   # 冻结原因 (9个值)\n    \"conts_type_509L\",          # 联系人类型 (9个值)\n    \"credacc_cards_status_52L\"  # 卡状态 (6个值，缺失97.5%)\n]\n\nprint(f\"✅ 配置完成: 需要对 {len(LOW_CARD_CAT_COLS)} 个类别字段进行透视展开。\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:23.550551Z","iopub.execute_input":"2026-08-23T10:26:23.550963Z","iopub.status.idle":"2026-08-23T10:26:23.567239Z","shell.execute_reply.started":"2026-08-23T10:26:23.550919Z","shell.execute_reply":"2026-08-23T10:26:23.566033Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: applprev_2 单表聚合函数 (修复 Null 广播问题)\n# ============================================================\n\ndef build_applprev_2_features(paths: list, verbose: bool = True) -> pl.DataFrame:\n    if verbose:\n        print(f\"[Step 1] 读取数据并采用宽松合并...\")\n    df = pl.concat([pl.scan_parquet(p) for p in paths], how=\"vertical_relaxed\")\n    \n    if verbose:\n        print(f\"[Step 2] 主键类型降级与显式类型转换...\")\n    # 🌟 修复点 1：显式将分类列强转为 String，防止由于全部空值被推断为纯 Null 导致聚合报错\n    df = df.with_columns([\n        pl.col(\"case_id\").cast(pl.UInt32),\n        *[pl.col(c).cast(pl.String) for c in LOW_CARD_CAT_COLS]\n    ])\n    \n    if verbose:\n        print(f\"[Step 3] 计算基础全局统计...\")\n    agg_exprs = [\n        # 🌟 修复点 2：不用 pl.len()，改用 pl.col(\"case_id\").count()\n        pl.col(\"case_id\").count().cast(pl.UInt32).alias(\"applprev_2_total_records\") \n    ]\n    for col in LOW_CARD_CAT_COLS:\n        agg_exprs.append(pl.col(col).n_unique().cast(pl.UInt8).alias(f\"{col}_nunique\"))\n        \n    main_agg = df.group_by(\"case_id\").agg(agg_exprs)\n    \n    if verbose:\n        print(f\"[Step 4] 配置低基数类别透视 (Count/Pct)...\")\n    cat_aggs = []\n    for col in LOW_CARD_CAT_COLS:\n        # 🌟 修复点 3：用 drop_nulls 替代 filter，用 col(\"case_id\").count() 替代 pl.len()\n        pivot_expr = (\n            df.drop_nulls(col)\n              .group_by([\"case_id\", col])\n              .agg(pl.col(\"case_id\").count().alias(\"cnt\")) \n        )\n        cat_aggs.append((col, pivot_expr))\n        \n    if verbose:\n        print(f\"[Step 5] 触发计算引擎 (engine='streaming')...\")\n    result = main_agg.collect(engine=\"streaming\")\n    \n    if verbose:\n        print(f\"[Step 6] 展开类别并合并宽表...\")\n    for col, cat_lazy in cat_aggs:\n        cat_df = cat_lazy.collect(engine=\"streaming\")\n        if cat_df.height == 0:\n            continue\n            \n        pivoted = cat_df.pivot(\n            values=\"cnt\",\n            index=\"case_id\",\n            on=col,\n            aggregate_function=\"sum\",\n        ).fill_null(0)\n        \n        new_cols = [c for c in pivoted.columns if c != \"case_id\"]\n        pivoted = pivoted.rename({c: f\"{col}_{c}_cnt\" for c in new_cols})\n        \n        cnt_cols = [f\"{col}_{c}_cnt\" for c in new_cols]\n        pivoted = pivoted.with_columns(pl.sum_horizontal(cnt_cols).alias(f\"{col}_total\"))\n        for c in new_cols:\n            pivoted = pivoted.with_columns(\n                (pl.col(f\"{col}_{c}_cnt\") / pl.col(f\"{col}_total\")).alias(f\"{col}_{c}_pct\")\n            )\n        pivoted = pivoted.drop(f\"{col}_total\")\n        \n        result = result.join(pivoted, on=\"case_id\", how=\"left\")\n        \n    if verbose:\n        print(f\"✅ applprev_2 清洗完毕，输出维度: {result.shape}\")\n    return result","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:23.568889Z","iopub.execute_input":"2026-08-23T10:26:23.569288Z","iopub.status.idle":"2026-08-23T10:26:23.588137Z","shell.execute_reply.started":"2026-08-23T10:26:23.569255Z","shell.execute_reply":"2026-08-23T10:26:23.587030Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 4: Phase A & B - 运行 Train 全量 (14M行建议直接全量，因为逻辑很简单)\n# ============================================================\nprint(\">>> 开始处理 Train applprev_2 (约1400万行) ...\")\nt0 = time.time()\ntrain_features = build_applprev_2_features(TRAIN_APPLPREV_2_PATHS)\nprint(f\"耗时: {time.time()-t0:.1f} 秒 | 内存: {train_features.estimated_size('mb'):.1f} MB\")\n\n\ntrain_features.write_parquet(OUT_TRAIN_PATH, compression=\"zstd\", compression_level=3)\ntrain_cols_list = train_features.columns\ndel train_features\ngc.collect()\n\n# ============================================================\n# Cell 5: Phase C - 运行 Test\n# ============================================================\nprint(\"\\n>>> 开始处理 Test applprev_2 ...\")\ntest_features = build_applprev_2_features(TEST_APPLPREV_2_PATHS)\n\n# 🌟 新增:强制对齐到 train schema\ntest_features = align_test_to_train(test_features, OUT_TRAIN_PATH)\n\ntest_features.write_parquet(OUT_TEST_PATH, compression=\"zstd\", compression_level=3)\ntest_cols_list = test_features.columns\n\nmissing_in_test = set(train_cols_list) - set(test_cols_list)\nextra_in_test = set(test_cols_list) - set(train_cols_list)\n\ndel test_features\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:23.589396Z","iopub.execute_input":"2026-08-23T10:26:23.589861Z","iopub.status.idle":"2026-08-23T10:26:30.713160Z","shell.execute_reply.started":"2026-08-23T10:26:23.589825Z","shell.execute_reply":"2026-08-23T10:26:30.711691Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6: 写日志与清理缓存\n# ============================================================\nlog_content = f\"\"\"============================================================\n数据清洗日志: applprev_2\n处理时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n\n[1. 字段丢弃]\n主动丢弃的无效/高缺失掩码字段 (0 个): 无\n\n[2. 训练集与测试集形态]\nTrain 特征列数 : {len(train_cols_list)} 列\nTest  特征列数 : {len(test_cols_list)} 列\n\n[3. Schema 差异 (Test 集未出现的低基数类别导致)]\nTrain 有但 Test 没有的列 ({len(missing_in_test)} 个):\n{', '.join(list(missing_in_test)) if missing_in_test else '无'}\n\nTest 有但 Train 没有的列 ({len(extra_in_test)} 个):\n{', '.join(list(extra_in_test)) if extra_in_test else '无'}\n============================================================\n\"\"\"\nprint(log_content)\n\nlog_path = OUT_ROOT / \"applprev_2_clean_log.txt\"\nwith open(log_path, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\n    \ngc.collect()\nprint(f\"🎉 applprev_2 完美收官！日志已保存至: {log_path}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:30.714412Z","iopub.execute_input":"2026-08-23T10:26:30.714751Z","iopub.status.idle":"2026-08-23T10:26:30.795093Z","shell.execute_reply.started":"2026-08-23T10:26:30.714721Z","shell.execute_reply":"2026-08-23T10:26:30.793714Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# credit_bureau_a_1 表\n\n（1）不同表结构不同：credit_bureau_a_1, credit_bureau_a_2, credit_bureau_b_1, credit_bureau_b_2, \n\nCredit Bureau A (征信局 A) 和 Credit Bureau B (征信局 B) 是两家完全不同的外部征信机构。\n\n由于机构不同，它们提供的数据维度、字段定义（Column 名字）以及覆盖的客户群（Coverage）是存在差异的。\n\n结论：A 和 B 必须完全独立处理，不能纵向合并（Concat）\n# -------------------------------------------------------\n\n（2）credit_bureau_a_1 和 credit_bureau_a_2 绝对不是同一个东西，它们是主表与明细表的关系（即 Depth=1 和 Depth=2 的关系）：\n\ncredit_bureau_a_1 (Depth=1, 贷款粒度):\n\n记录的是客户在征信局的历史贷款主记录\n\n主键是 case_id + num_group1\n\n举例：客户（case_id=100）过去在外部借了 3 笔钱，那么在 A_1 表里会有 3 行记录（num_group1 为 0, 1, 2），包含这 3 笔贷款的额度、状态等。\n\ncredit_bureau_a_2 (Depth=2, 流水粒度):\n\n记录的是对应历史贷款的底层流水明细（通常是每期的还款记录、逾期天数等）。\n\n主键是 case_id + num_group1 + num_group2。\n\n举例：客户的那 3 笔贷款，第一笔分了 12 期，第二笔分了 6 期，第三笔分了 24 期。那么在 A_2 表里，就会有 12 + 6 + 24 = 42 行记录，用 num_group2 来标记期数。\n# -------------------------------------------------------\n\n（3）表有以下几个极其重要的特征：\n\n数据体量巨大：分成 4 个分片，共约 1600 万行（比 applprev_1 还要庞大两倍多）。\n\n极度稀疏：高达 79 列特征中，有大量的字段缺失率超过 90% 甚至 99%（比如 prolongationcount_599L, interestrate_508L 等）。\n\n常数值（Near Constant）泛滥：许多金额类或逾期类的字段，99% 的值都是 0，或者某些掩码字符串只有 a55475b1 这一个高频值。\n\n日期特征繁杂：包含了各种 active/closed 合同的开始、结束、逾期最大月份等 21 个与时间相关的字段。\n\n由于这依然是一张 Depth=1 的表（每个 case_id 有多条历史合同，平均约 11.5 条），我们的处理逻辑和之前的 applprev_1 高度一致：平行拍平法，一步降维至 case_id 级别。\n# -------------------------------------------------------\n\n（4）删除字段\n\n判断一个特征是否属于“绝对垃圾”，核心标准是：它能否为模型提供区分正负样本（违约 vs 不违约）的“信息增益”（Information Gain）\n\n以下是这三个字段必须被删除的绝对理由（从底层算法和数据科学角度剖析）：\n\n1. description_351M (类别特征)\n\n数据表现： EDA 显示该列 100% 的值都是掩码字符串 'a55475b1'（或者少量的纯空值 Null，在类别特征里 Null 通常也会被当成一类）。\n\n绝对删除理由：零方差（Zero Variance）。\n\n如果所有客户在这个特征上的取值都一模一样，树模型在构建决策树时，根本无法根据这个特征进行分裂（Split）。\n\n因为无论怎么切一刀，左右两边的样本在这个特征上都是一样的，算出来的信息增益绝对等于 0。\n\n结论： 纯粹占用内存，对预测贡献为 0\n\n2. prolongationcount_599L (数值特征，合同展期次数)\n\n数据表现： 缺失率高达 99.7%，且剩下的那 0.3% 非空数据中，绝大多数（甚至全部）的值都是 0.0。\n\n绝对删除理由：极度稀疏 + 近乎零方差。\n\n无法泛化： 只有 0.3% 的人有数据，意味着在一万个人里只有 30 个人有记录。即使这 30 个人里刚好有几个人违约了，模型如果抓取了这个特征，极大概率是过拟合（Overfitting）到了这极少数的样本上。\n\n把空值填为 0，那整个列 100% 都是 0；如果你把空值填为 -1，那模型只能学到“有记录（值为0）”和“没记录（值为-1）”的区别。但在绝大多数非空值也是 0 的情况下，这种区分几乎没有任何业务意义。\n\n结论：引入该特征极易导致过拟合，且不包含有效的区分信息。\n\n\n3. interestrate_508L (数值特征，历史闭环合同利率)\n\n数据表现： 缺失率高达 99.5%。\n\n绝对删除理由：严重稀疏导致的稳定性风险（严重违背本比赛核心考点）。\n\n这场 Kaggle 比赛的核心目标是 Model Stability（模型稳定性），评估指标 Gini Stability 会严惩随着时间推移性能下降的模型。\n\n一个缺失率 99.5% 的利率字段，往往是因为某个历史特定时期（比如仅在 2017 年某几个月的某项特定业务中）才被采集过。\n\n如果你的模型在训练集里勉强用这 0.5% 的数据学到了一点点微弱的规律，一旦到了测试集（未来的时间段），由于业务规则改变，这个字段可能彻底变为 100% 缺失，会导致模型在这一维度的预测逻辑直接崩塌，从而大幅拉低模型的长期稳定性得分。\n\n结论： 极度稀疏的业务数值特征是破坏模型时间稳定性的定时炸弹。\n\n保留这三个字段，不仅无法提供任何有效的信息增益，还会白白消耗你进行 group_by 聚合时的 CPU 算力和宝贵的内存（Kaggle 只有 16GB/30GB 内存限制），并且 interestrate_508L 还会增加模型在未来测试集上表现不稳定的风险。\n# -------------------------------------------------------\n\n（5）处理策略主动丢弃：\n\n仅丢弃那些“缺失率 $\\ge$ 99% 且方差极小/近乎常数”或者“完全是单一无意义掩码”的字段。\n\nEDA 探查，description_351M（全样本几乎都是 'a55475b1'）、prolongationcount_599L（缺失 99.7% 且全是 0）、interestrate_508L（缺失 99.5%）这些属于绝对垃圾，我们将其扔掉。\n\n极度稀疏的数值处理：对于那些缺失率 90% 以上但方差不为 0 的数值字段（如各种逾期金额），我们保留，因为“发生过严重逾期”这种罕见事件在风控里是强信号。\n\n聚合时，除了求 sum/max，它们的“非空计数 (cnt)”本身就是一个好特征。日期字段：不与 base 表关联，直接在单表内求最近（max）和最远（min）日期。\n\n架构：延续上个表的优雅结构，包含完整的日志输出 ","metadata":{}},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 路径与环境准备 (credit_bureau_a_1)\n# ============================================================\nimport polars as pl\nimport gc\nimport time\nfrom pathlib import Path\nfrom datetime import datetime\n\n# 1. 定义数据源根目录 (Kaggle 环境)\nDATA_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability/parquet_files\")\nTRAIN_DIR = DATA_ROOT / \"train\"\nTEST_DIR = DATA_ROOT / \"test\"\n\n# 2. 获取所有分片路径 (共 4 个 train 分片)\nTRAIN_CBA1_PATHS = sorted(list(TRAIN_DIR.glob(\"train_credit_bureau_a_1_*.parquet\")))\nTEST_CBA1_PATHS = sorted(list(TEST_DIR.glob(\"test_credit_bureau_a_1_*.parquet\")))\n\n# 3. 定义输出目录\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# 4. 定义聚合后文件的保存路径\nOUT_TRAIN_PATH = OUT_ROOT / \"credit_bureau_a_1_clean_train.parquet\"\nOUT_TEST_PATH = OUT_ROOT / \"credit_bureau_a_1_clean_test.parquet\"\n\nprint(f\"✅ 环境准备完毕:\")\nprint(f\"   Train 分片数: {len(TRAIN_CBA1_PATHS)}\")\nprint(f\"   Test  分片数: {len(TEST_CBA1_PATHS)}\")\nassert len(TRAIN_CBA1_PATHS) > 0, \"❌ 未找到 train_credit_bureau_a_1\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:30.796463Z","iopub.execute_input":"2026-08-23T10:26:30.796828Z","iopub.status.idle":"2026-08-23T10:26:30.808471Z","shell.execute_reply.started":"2026-08-23T10:26:30.796797Z","shell.execute_reply":"2026-08-23T10:26:30.807273Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: 字段分类配置 (基于 credit_bureau_a_1 的探查结果)\n# ============================================================\n\n# --- 主键 ---\nKEY_COLS = [\"case_id\", \"num_group1\"]\n\n# --- 绝对垃圾字段 (缺失 >99.5% 或完全单一无意义) ---\nDROP_COLS = [\n    \"description_351M\",        # 100% 都是 'a55475b1'\n    \"prolongationcount_599L\",  # 缺失 99.7%，非空也是 0\n    \"interestrate_508L\",       # 缺失 99.5%\n]\n\n# --- 类别特征 (做 n_unique) ---\nCAT_COLS = [\n    \"classificationofcontr_13M\", \"classificationofcontr_400M\",\n    \"contractst_545M\", \"contractst_964M\",\n    \"financialinstitution_382M\", \"financialinstitution_591M\",\n    \"purposeofcred_426M\", \"purposeofcred_874M\",\n    \"subjectrole_182M\", \"subjectrole_93M\"\n]\n\n# --- 日期特征 (求极值 max/min) ---\nDATE_COLS = [\n    \"dateofcredend_289D\", \"dateofcredend_353D\", \"dateofcredstart_181D\",\n    \"dateofcredstart_739D\", \"dateofrealrepmt_138D\", \"lastupdate_1112D\",\n    \"lastupdate_388D\", \"numberofoverdueinstlmaxdat_148D\", \n    \"numberofoverdueinstlmaxdat_641D\", \"overdueamountmax2date_1002D\", \n    \"overdueamountmax2date_1142D\", \"refreshdate_3813885D\"\n]\n\n# --- 其余全部视为数值特征 (求 max/mean/sum/cnt 等) ---\n# 注意：年份/月份结尾为 T 的特征，我们也当数值算 max/min 即可。\nALL_KNOWN = set(KEY_COLS + DROP_COLS + CAT_COLS + DATE_COLS)\n# (将在读取单分片时自动推断剩余的数值列，省去手敲几十个字段)\n\nprint(f\"✅ 字段分类策略配置完成。\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:30.810061Z","iopub.execute_input":"2026-08-23T10:26:30.810540Z","iopub.status.idle":"2026-08-23T10:26:30.830114Z","shell.execute_reply.started":"2026-08-23T10:26:30.810492Z","shell.execute_reply":"2026-08-23T10:26:30.828775Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: 核心单表聚合函数\n# ============================================================\n\ndef build_cba1_features(paths: list, verbose: bool = True) -> pl.DataFrame:\n    if verbose:\n        print(f\"[Step 1] 读取数据并采用宽松合并...\")\n    df = pl.concat([pl.scan_parquet(p) for p in paths], how=\"vertical_relaxed\")\n    \n    if verbose:\n        print(f\"[Step 2] 丢弃垃圾字段并转换基础类型...\")\n    df = df.drop(DROP_COLS)\n    \n    df = df.with_columns(pl.col(\"case_id\").cast(pl.UInt32))\n    for c in DATE_COLS:\n        df = df.with_columns(pl.col(c).str.to_date(\"%Y-%m-%d\", strict=False))\n    \n    # 动态获取所有数值列 (排除掉键、日期、类别和已丢弃的)\n    current_cols = set(df.columns)\n    NUMERIC_COLS = list(current_cols - set(KEY_COLS + CAT_COLS + DATE_COLS))\n    \n    if verbose:\n        print(f\"   -> 动态识别出 {len(NUMERIC_COLS)} 个数值/浮点/T后缀列。\")\n        print(f\"[Step 3] 配置全局聚合算子...\")\n        \n    agg_exprs = [pl.col(\"case_id\").count().cast(pl.UInt16).alias(\"cba1_total_records\")]\n    \n    # 1. 类别列 -> count 独立类别数\n    for c in CAT_COLS:\n        agg_exprs.append(pl.col(c).n_unique().cast(pl.UInt8).alias(f\"{c}_nunique\"))\n        \n    # 2. 日期列 -> min (最早) / max (最晚)\n    for c in DATE_COLS:\n        agg_exprs.extend([\n            pl.col(c).min().alias(f\"{c}_min\"),\n            pl.col(c).max().alias(f\"{c}_max\"),\n        ])\n        \n    # 3. 数值列 -> sum (总和) / max (极值，比如最大逾期) / mean (平均) / cnt (非空次数)\n    for c in NUMERIC_COLS:\n        agg_exprs.extend([\n            pl.col(c).max().alias(f\"{c}_max\"),\n            pl.col(c).mean().alias(f\"{c}_mean\"),\n            pl.col(c).sum().alias(f\"{c}_sum\"),\n            pl.col(c).is_not_null().sum().cast(pl.UInt16).alias(f\"{c}_notnull_cnt\"),\n        ])\n        \n    if verbose:\n        print(f\"[Step 4] 执行 group_by 并触发 streaming 计算 (耗时较长)...\")\n        \n    # 因为没有复杂的 pivot (透视)，这表直接 group_by 就结束了，对内存非常友好！\n    result = df.group_by(\"case_id\").agg(agg_exprs).collect(engine=\"streaming\")\n    \n    if verbose:\n        print(f\"✅ credit_bureau_a_1 清洗完毕，输出维度: {result.shape}\")\n        \n    return result","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:30.832101Z","iopub.execute_input":"2026-08-23T10:26:30.832821Z","iopub.status.idle":"2026-08-23T10:26:30.859412Z","shell.execute_reply.started":"2026-08-23T10:26:30.832761Z","shell.execute_reply":"2026-08-23T10:26:30.858234Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 4: 运行 Train 全量 (约 1600 万行)\n# ============================================================\nprint(\">>> 开始处理 Train credit_bureau_a_1 ...\")\nt0 = time.time()\ntrain_features = build_cba1_features(TRAIN_CBA1_PATHS)\nprint(f\"耗时: {time.time()-t0:.1f} 秒 | 内存: {train_features.estimated_size('mb'):.1f} MB\")\n\ntrain_features.write_parquet(OUT_TRAIN_PATH, compression=\"zstd\", compression_level=3)\ntrain_cols_list = train_features.columns\ndel train_features\ngc.collect()\n\n# ============================================================\n# Cell 5: 运行 Test\n# ============================================================\nprint(\"\\n>>> 开始处理 Test credit_bureau_a_1 ...\")\ntest_features = build_cba1_features(TEST_CBA1_PATHS)\ntest_features.write_parquet(OUT_TEST_PATH, compression=\"zstd\", compression_level=3)\ntest_cols_list = test_features.columns\n\nmissing_in_test = set(train_cols_list) - set(test_cols_list)\nextra_in_test = set(test_cols_list) - set(train_cols_list)\n\ndel test_features\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:26:30.860853Z","iopub.execute_input":"2026-08-23T10:26:30.861187Z","iopub.status.idle":"2026-08-23T10:27:28.952369Z","shell.execute_reply.started":"2026-08-23T10:26:30.861154Z","shell.execute_reply":"2026-08-23T10:27:28.951415Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6: 写日志与清理缓存\n# ============================================================\nlog_content = f\"\"\"============================================================\n数据清洗日志: credit_bureau_a_1\n处理时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n\n[1. 字段丢弃]\n主动丢弃的无效/极度稀疏掩码字段 ({len(DROP_COLS)} 个):\n{', '.join(DROP_COLS)}\n\n[2. 训练集与测试集形态]\nTrain 特征列数 : {len(train_cols_list)} 列\nTest  特征列数 : {len(test_cols_list)} 列\n\n[3. Schema 差异 (因为本次未使用 Pivot 展开类别，差异应为 0)]\nTrain 有但 Test 没有的列 ({len(missing_in_test)} 个): \n{', '.join(list(missing_in_test)) if missing_in_test else '无'}\n\nTest 有但 Train 没有的列 ({len(extra_in_test)} 个):\n{', '.join(list(extra_in_test)) if extra_in_test else '无'}\n============================================================\n\"\"\"\nprint(log_content)\n\nlog_path = OUT_ROOT / \"credit_bureau_a_1_clean_log.txt\"\nwith open(log_path, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\n    \ngc.collect()\ngc.collect()\nprint(f\"🎉 credit_bureau_a_1 处理完美收官！日志已保存至: {log_path}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:27:28.953850Z","iopub.execute_input":"2026-08-23T10:27:28.954195Z","iopub.status.idle":"2026-08-23T10:27:29.109364Z","shell.execute_reply.started":"2026-08-23T10:27:28.954165Z","shell.execute_reply":"2026-08-23T10:27:29.108276Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# credit_bureau_a_2（征信局 A 提供的历史贷款底层还款流水表）\n\n（1）此表绝对是本次比赛中的“重头戏”。\n\n这表的特征太鲜明了：\n\n数据量极其恐怖：拆成了 11 个分片，总行数接近 1.88 亿行（188,298,452）！这是典型的 Depth=2 表，平均每个客户有 135.9 条流水记录。\n\n纯粹的流水信息：全表只有 19 列，除了主键，全是逾期天数（DPD）、还款年月、逾期金额以及抵押物信息。\n# --------------------------------------------------------------------------------\n\n（2）垃圾字段判定：\n\n缺失 >= 99% 且极低方差，或 100% 纯掩码）”的死刑标准。我仔细审查了你的 EDA 报告，结论是：在 credit_bureau_a_2 中，没有任何一个字段达到了“必须被枪毙”的绝对标准，全部予以保留！\n\n不杀的绝对理由：\n> 1.缺失率未触及红线：全表缺失率最高的是 collater_valueofguarantee_1124L (98.5%) 和 collater_valueofguarantee_876L (96.3%)。在风控中，抵押物数据通常就非常稀少，但这 1.5%；3.7% 的有抵押数据往往对降低风险预测有着决定性作用。不能因为少就乱删；\n>\n> 2.分类字段包含有效信息：像 collater_typofvalofguarant_298M 和 subjectroles_name_541M 虽然 'a55475b1' 占了绝大多数，但它们依然有 'ab3c25cf'、'8fd95e4b' 等数十甚至成百上千条独立分类，这意味着它们并非零方差掩码，而是严重的“长尾分布”。我们通过计算 nunique 即可提炼出“业务复杂度”特征。\n>\n> 3.逾期流水是风控的灵魂：pmts_dpd（逾期天数）和 pmts_overdue（逾期金额）相关字段虽然含有大量的 0（按时还款），但这恰恰是判断信用好坏的基石，极其宝贵。\n>\n> 4.针对这 1.88 亿行的怪兽级数据，我们继续采用“平行拍平法”，无视 num_group1 和 num_group2，直接把它们砸平成每个客户一行\n# --------------------------------------------------------------------------------\n\n\n（3）\n","metadata":{}},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 路径与环境准备 (credit_bureau_a_2)\n# ============================================================\nimport polars as pl\nimport gc\nimport time\nfrom pathlib import Path\nfrom datetime import datetime\n\n# 1. 定义数据源根目录\nDATA_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability/parquet_files\")\nTRAIN_DIR = DATA_ROOT / \"train\"\nTEST_DIR = DATA_ROOT / \"test\"\n\n# 2. 获取所有分片路径 (共 11 个 train 分片，约 1.88 亿行！)\nTRAIN_CBA2_PATHS = sorted(list(TRAIN_DIR.glob(\"train_credit_bureau_a_2_*.parquet\")))\nTEST_CBA2_PATHS = sorted(list(TEST_DIR.glob(\"test_credit_bureau_a_2_*.parquet\")))\n\n# 3. 定义输出目录\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# 4. 定义聚合后文件的保存路径\nOUT_TRAIN_PATH = OUT_ROOT / \"credit_bureau_a_2_clean_train.parquet\"\nOUT_TEST_PATH = OUT_ROOT / \"credit_bureau_a_2_clean_test.parquet\"\n\nprint(f\"✅ 环境准备完毕:\")\nprint(f\"   Train 分片数: {len(TRAIN_CBA2_PATHS)}\")\nprint(f\"   Test  分片数: {len(TEST_CBA2_PATHS)}\")\nassert len(TRAIN_CBA2_PATHS) > 0, \"❌ 未找到 train_credit_bureau_a_2\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:27:29.110644Z","iopub.execute_input":"2026-08-23T10:27:29.111027Z","iopub.status.idle":"2026-08-23T10:27:29.138012Z","shell.execute_reply.started":"2026-08-23T10:27:29.110981Z","shell.execute_reply":"2026-08-23T10:27:29.136696Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: 字段分类配置 (基于 credit_bureau_a_2 探查)\n# ============================================================\n\n# --- 主键 ---\nKEY_COLS = [\"case_id\", \"num_group1\", \"num_group2\"]\n\n# --- 绝对垃圾字段 (经查证：无) ---\nDROP_COLS = []\n\n# --- 类别特征 (计算 n_unique 复杂度) ---\nCAT_COLS = [\n    \"collater_typofvalofguarant_298M\", \n    \"collater_typofvalofguarant_407M\",\n    \"collaterals_typeofguarante_359M\", \n    \"collaterals_typeofguarante_669M\",\n    \"subjectroles_name_541M\", \n    \"subjectroles_name_838M\"\n]\n\n# --- 特殊风控字段：逾期天数 (提取严重逾期次数) ---\nDPD_COLS = [\n    \"pmts_dpd_1073P\", \n    \"pmts_dpd_303P\"\n]\n\n# (将在代码内自动推断剩余的所有数值类特征，如 amount, year, month)\nprint(f\"✅ 字段分类配置完毕：保留全量特征，将压平这 1.88 亿行数据。\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:27:29.139320Z","iopub.execute_input":"2026-08-23T10:27:29.139999Z","iopub.status.idle":"2026-08-23T10:27:29.157697Z","shell.execute_reply.started":"2026-08-23T10:27:29.139964Z","shell.execute_reply":"2026-08-23T10:27:29.156616Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: 核心单表聚合函数\n# ============================================================\n\ndef build_cba2_features(paths: list, verbose: bool = True) -> pl.DataFrame:\n    if verbose:\n        print(f\"[Step 1] 读取 {len(paths)} 个分片并采用宽松合并...\")\n    df = pl.concat([pl.scan_parquet(p) for p in paths], how=\"vertical_relaxed\")\n    \n    if verbose:\n        print(f\"[Step 2] 显式类型转换 (防空洞/内存优化)...\")\n    # 强制类别字段为 String 防止空分片推断错误\n    df = df.with_columns([\n        pl.col(\"case_id\").cast(pl.UInt32),\n        *[pl.col(c).cast(pl.String) for c in CAT_COLS]\n    ])\n    \n    # 获取数值列\n    current_cols = set(df.columns)\n    NUMERIC_COLS = list(current_cols - set(KEY_COLS + CAT_COLS + DROP_COLS))\n    \n    if verbose:\n        print(f\"[Step 3] 配置全局聚合算子...\")\n    \n    # 基础：这客户总共有多少条流水\n    agg_exprs = [pl.col(\"case_id\").count().cast(pl.UInt32).alias(\"cba2_total_payments\")]\n    \n    # 类别列：有几种不同类型\n    for c in CAT_COLS:\n        agg_exprs.append(pl.col(c).n_unique().cast(pl.UInt8).alias(f\"{c}_nunique\"))\n        \n    # 数值列：max/mean/sum/cnt\n    for c in NUMERIC_COLS:\n        agg_exprs.extend([\n            pl.col(c).max().alias(f\"{c}_max\"),\n            pl.col(c).mean().alias(f\"{c}_mean\"),\n            pl.col(c).sum().alias(f\"{c}_sum\"),\n            pl.col(c).is_not_null().sum().cast(pl.UInt32).alias(f\"{c}_notnull_cnt\"),\n        ])\n        \n    # 特别风控逻辑：有过任何逾期的流水次数 (DPD > 0)\n    for c in DPD_COLS:\n        agg_exprs.append(\n            pl.col(c).filter(pl.col(c) > 0).len().cast(pl.UInt32).alias(f\"{c}_gt0_cnt\")\n        )\n\n    if verbose:\n        print(f\"[Step 4] 触发 streaming 引擎进行 1.88 亿行的终极聚合...\")\n    \n    # 因为数据量太大，完全依赖 streaming 分批运算\n    result = df.group_by(\"case_id\").agg(agg_exprs).collect(engine=\"streaming\")\n    \n    if verbose:\n        print(f\"✅ credit_bureau_a_2 清洗完毕，输出维度: {result.shape}\")\n        \n    return result","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:27:29.158995Z","iopub.execute_input":"2026-08-23T10:27:29.159375Z","iopub.status.idle":"2026-08-23T10:27:29.187989Z","shell.execute_reply.started":"2026-08-23T10:27:29.159332Z","shell.execute_reply":"2026-08-23T10:27:29.186698Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: 核心单表聚合函数 (防 OOM 分块处理版)\n\n# 我们必须采用大数据处理中最经典的 “分块处理（Map-Reduce）” 策略：\n# 我们不再一上来就把 11 个文件拼在一起，而是写一个 for 循环，每次只把 1 个文件读进内存，做完 group_by 聚合后（几千万行瞬间被压缩成几万行），\n# 再处理下一个文件。最后把这 11 个高度压缩后的“精华”拼起来即可。\n# ============================================================\n\ndef build_cba2_features(paths: list, verbose: bool = True) -> pl.DataFrame:\n    chunks = []\n    \n    # 🌟 [核心优化] 不再一上来就 concat，而是逐个文件单独突破！\n    for i, p in enumerate(paths):\n        if verbose:\n            print(f\"   -> [分块 {i+1}/{len(paths)}] 正在读取并处理: {p.name} ...\")\n            \n        # 1. 直接全量读取单文件 (单文件只有几十 MB，放进内存毫无压力)\n        df = pl.read_parquet(p)\n        \n        # 2. 显式类型转换\n        df = df.with_columns([\n            pl.col(\"case_id\").cast(pl.UInt32),\n            *[pl.col(c).cast(pl.String) for c in CAT_COLS]\n        ])\n        \n        current_cols = set(df.columns)\n        NUMERIC_COLS = list(current_cols - set(KEY_COLS + CAT_COLS + DROP_COLS))\n        \n        # 3. 配置聚合算子\n        agg_exprs = [pl.col(\"case_id\").count().cast(pl.UInt32).alias(\"cba2_total_payments\")]\n        \n        for c in CAT_COLS:\n            agg_exprs.append(pl.col(c).n_unique().cast(pl.UInt8).alias(f\"{c}_nunique\"))\n            \n        for c in NUMERIC_COLS:\n            agg_exprs.extend([\n                pl.col(c).max().alias(f\"{c}_max\"),\n                pl.col(c).mean().alias(f\"{c}_mean\"),\n                pl.col(c).sum().alias(f\"{c}_sum\"),\n                pl.col(c).is_not_null().sum().cast(pl.UInt32).alias(f\"{c}_notnull_cnt\"),\n            ])\n            \n        for c in DPD_COLS:\n            # 优化：用布尔值 sum 替代 filter，极大提升计算速度\n            agg_exprs.append(\n                (pl.col(c) > 0).sum().cast(pl.UInt32).alias(f\"{c}_gt0_cnt\")\n            )\n            \n        # 4. 执行单文件聚合 (把几千万行流水，瞬间压缩成几万行客户特征)\n        chunk_df = df.group_by(\"case_id\").agg(agg_exprs)\n        chunks.append(chunk_df)\n        \n        # 🌟 主动清理单文件的原始内存，释放 RAM\n        del df\n        gc.collect()\n\n    if verbose:\n        print(\"\\n[Step Final] 合并所有分块结果，处理边缘 case_id...\")\n        \n    # 5. 合并所有高度压缩后的块 (11 个小 DataFrame 合并，内存占用不到 100MB)\n    result = pl.concat(chunks, how=\"vertical_relaxed\")\n    \n    # 6. 去重跨界 case_id\n    # (Kaggle 源数据是按 case_id 排序切分的，仅有极少数 case_id 会刚好跨越两个文件，直接保留第一段的特征即可)\n    result = result.unique(subset=[\"case_id\"], keep=\"first\")\n    \n    if verbose:\n        print(f\"✅ credit_bureau_a_2 清洗完毕，输出维度: {result.shape}\")\n        \n    return result","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:27:29.189432Z","iopub.execute_input":"2026-08-23T10:27:29.189815Z","iopub.status.idle":"2026-08-23T10:27:29.213495Z","shell.execute_reply.started":"2026-08-23T10:27:29.189776Z","shell.execute_reply":"2026-08-23T10:27:29.212211Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 4: 运行 Train 全量 (请耐心等待，约耗时 1~3 分钟)\n# ============================================================\nprint(\">>> 开始处理 Train credit_bureau_a_2 (这是硬仗，请耐心)...\")\nt0 = time.time()\ntrain_features = build_cba2_features(TRAIN_CBA2_PATHS)\nprint(f\"耗时: {time.time()-t0:.1f} 秒 | 内存: {train_features.estimated_size('mb'):.1f} MB\")\n\ntrain_features.write_parquet(OUT_TRAIN_PATH, compression=\"zstd\", compression_level=3)\ntrain_cols_list = train_features.columns\ndel train_features\ngc.collect()\n\n# ============================================================\n# Cell 5: 运行 Test\n# ============================================================\nprint(\"\\n>>> 开始处理 Test credit_bureau_a_2 ...\")\ntest_features = build_cba2_features(TEST_CBA2_PATHS)\ntest_features.write_parquet(OUT_TEST_PATH, compression=\"zstd\", compression_level=3)\ntest_cols_list = test_features.columns\n\nmissing_in_test = set(train_cols_list) - set(test_cols_list)\nextra_in_test = set(test_cols_list) - set(train_cols_list)\n\ndel test_features\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:27:29.214961Z","iopub.execute_input":"2026-08-23T10:27:29.215327Z","iopub.status.idle":"2026-08-23T10:28:12.681643Z","shell.execute_reply.started":"2026-08-23T10:27:29.215294Z","shell.execute_reply":"2026-08-23T10:28:12.680656Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6: 写日志与清理缓存\n# ============================================================\nlog_content = f\"\"\"============================================================\n数据清洗日志: credit_bureau_a_2\n处理时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n\n[1. 字段丢弃]\n主动丢弃的无效字段 ({len(DROP_COLS)} 个):\n{', '.join(DROP_COLS) if DROP_COLS else '无 (本表无绝对垃圾，全数保留)'}\n\n[2. 训练集与测试集形态]\nTrain 特征列数 : {len(train_cols_list)} 列\nTest  特征列数 : {len(test_cols_list)} 列\n\n[3. Schema 差异 (因为本次未使用 Pivot 展开类别，差异应为 0)]\nTrain 有但 Test 没有的列 ({len(missing_in_test)} 个):\n{', '.join(list(missing_in_test)) if missing_in_test else '无'}\n\nTest 有但 Train 没有的列 ({len(extra_in_test)} 个):\n{', '.join(list(extra_in_test)) if extra_in_test else '无'}\n============================================================\n\"\"\"\nprint(log_content)\n\nlog_path = OUT_ROOT / \"credit_bureau_a_2_clean_log.txt\"\nwith open(log_path, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\n    \ngc.collect()\ngc.collect()\nprint(f\"🎉 1.88 亿行的巨兽 credit_bureau_a_2 降服完毕！日志已保存至: {log_path}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:12.683025Z","iopub.execute_input":"2026-08-23T10:28:12.683434Z","iopub.status.idle":"2026-08-23T10:28:12.840045Z","shell.execute_reply.started":"2026-08-23T10:28:12.683392Z","shell.execute_reply":"2026-08-23T10:28:12.839000Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# credit_bureau_b_1 表（征信局 B 提供的历史借款主记录表）\n\n表的数据形态让人感到“如释重负”：\n\n体量极小：全量只有 1 个文件，8.5 万行，仅 4.3 MB。与刚才 1.88 亿行的 A 局流水表相比，简直是沧海一粟。\n\n深度为 1：平均每个客户在 B 局只有 2.35 条贷款记录。\n\n没有绝对垃圾字段：最高缺失率只有 81.2%（residualamount_1093A），而且没有任何缺失率达到 99% 的字段，所以根据我们的“死刑标准”，本表不需要丢弃任何列，全量保留！\n\n因为数据量极小，内存毫无压力，我们可以直接用最清爽的单次加载聚合。以下是全套清洗代码（延续了咱们之前的优雅风格与日志输出）：\n","metadata":{}},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 路径与环境准备 (credit_bureau_b_1)\n# ============================================================\nimport polars as pl\nimport gc\nimport time\nfrom pathlib import Path\nfrom datetime import datetime\n\n# 1. 定义数据源根目录 (Kaggle 环境)\nDATA_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability/parquet_files\")\nTRAIN_DIR = DATA_ROOT / \"train\"\nTEST_DIR = DATA_ROOT / \"test\"\n\n# 2. 获取所有分片路径 (自动匹配可能带后缀的格式)\nTRAIN_CBB1_PATHS = sorted(list(TRAIN_DIR.glob(\"train_credit_bureau_b_1*.parquet\")))\nTEST_CBB1_PATHS = sorted(list(TEST_DIR.glob(\"test_credit_bureau_b_1*.parquet\")))\n\n# 3. 定义输出目录\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# 4. 定义聚合后文件的保存路径\nOUT_TRAIN_PATH = OUT_ROOT / \"credit_bureau_b_1_clean_train.parquet\"\nOUT_TEST_PATH = OUT_ROOT / \"credit_bureau_b_1_clean_test.parquet\"\n\nprint(f\"✅ 环境准备完毕:\")\nprint(f\"   Train 分片数: {len(TRAIN_CBB1_PATHS)}\")\nprint(f\"   Test  分片数: {len(TEST_CBB1_PATHS)}\")\nassert len(TRAIN_CBB1_PATHS) > 0, \"❌ 未找到 train_credit_bureau_b_1\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:12.841377Z","iopub.execute_input":"2026-08-23T10:28:12.841804Z","iopub.status.idle":"2026-08-23T10:28:12.852546Z","shell.execute_reply.started":"2026-08-23T10:28:12.841761Z","shell.execute_reply":"2026-08-23T10:28:12.851617Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: 字段分类配置 (基于 credit_bureau_b_1 探查)\n# ============================================================\n\n# --- 主键 ---\nKEY_COLS = [\"case_id\", \"num_group1\"]\n\n# --- 绝对垃圾字段 (经评估，最高缺失仅 81%，无须丢弃) ---\nDROP_COLS = []\n\n# --- 类别特征 (做 n_unique) ---\nCAT_COLS = [\n    \"classificationofcontr_1114M\", \"contractst_516M\", \n    \"contracttype_653M\", \"credor_3940957M\", \n    \"periodicityofpmts_997L\", \"periodicityofpmts_997M\", \n    \"pmtmethod_731M\", \"purposeofcred_722M\", \n    \"subjectrole_326M\", \"subjectrole_43M\"\n]\n\n# --- 严格的日期特征 (求 max/min，需 to_date 转换) ---\nDATE_COLS = [\n    \"contractdate_551D\", \n    \"contractmaturitydate_151D\", \n    \"lastupdate_260D\"\n]\n\n# (说明：对于 T 结尾的年/月特征，如 dpdmaxdatemonth_804T，\n# 它本身就是 Float64，我们让代码自动将其视为数值，直接求 max/min/mean 即可)\n\nprint(f\"✅ 字段分类配置完成，将提取独立类别数、日期边界和各类金额的统计值。\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:12.854137Z","iopub.execute_input":"2026-08-23T10:28:12.854927Z","iopub.status.idle":"2026-08-23T10:28:12.873811Z","shell.execute_reply.started":"2026-08-23T10:28:12.854892Z","shell.execute_reply":"2026-08-23T10:28:12.872640Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: 核心单表聚合函数\n# ============================================================\n\ndef build_cbb1_features(paths: list, verbose: bool = True) -> pl.DataFrame:\n    if verbose:\n        print(f\"[Step 1] 读取数据并采用宽松合并...\")\n    df = pl.concat([pl.scan_parquet(p) for p in paths], how=\"vertical_relaxed\")\n    \n    if verbose:\n        print(f\"[Step 2] 转换基础类型 (主键与日期)...\")\n        \n    # 主键、类别显式转 String 防止空分片报错，日期列转换\n    df = df.with_columns([\n        pl.col(\"case_id\").cast(pl.UInt32),\n        *[pl.col(c).cast(pl.String) for c in CAT_COLS]\n    ])\n    \n    for c in DATE_COLS:\n        df = df.with_columns(pl.col(c).str.to_date(\"%Y-%m-%d\", strict=False))\n        \n    # 动态获取数值列\n    current_cols = set(df.columns)\n    NUMERIC_COLS = list(current_cols - set(KEY_COLS + CAT_COLS + DATE_COLS))\n    \n    if verbose:\n        print(f\"[Step 3] 配置全局聚合算子...\")\n        \n    agg_exprs = [pl.col(\"case_id\").count().cast(pl.UInt16).alias(\"cbb1_total_records\")]\n    \n    for c in CAT_COLS:\n        agg_exprs.append(pl.col(c).n_unique().cast(pl.UInt8).alias(f\"{c}_nunique\"))\n        \n    for c in DATE_COLS:\n        agg_exprs.extend([\n            pl.col(c).min().alias(f\"{c}_min\"),\n            pl.col(c).max().alias(f\"{c}_max\"),\n        ])\n        \n    for c in NUMERIC_COLS:\n        agg_exprs.extend([\n            pl.col(c).max().alias(f\"{c}_max\"),\n            pl.col(c).mean().alias(f\"{c}_mean\"),\n            pl.col(c).sum().alias(f\"{c}_sum\"),\n            pl.col(c).is_not_null().sum().cast(pl.UInt16).alias(f\"{c}_notnull_cnt\"),\n        ])\n        \n    if verbose:\n        print(f\"[Step 4] 执行聚合 (数据量小，瞬间完成)...\")\n        \n    result = df.group_by(\"case_id\").agg(agg_exprs).collect(engine=\"streaming\")\n    \n    if verbose:\n        print(f\"✅ credit_bureau_b_1 清洗完毕，输出维度: {result.shape}\")\n        \n    return result","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:12.875252Z","iopub.execute_input":"2026-08-23T10:28:12.875848Z","iopub.status.idle":"2026-08-23T10:28:12.896815Z","shell.execute_reply.started":"2026-08-23T10:28:12.875799Z","shell.execute_reply":"2026-08-23T10:28:12.895725Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 4: 运行 Train 全量 (仅 8.5 万行)\n# ============================================================\nprint(\">>> 开始处理 Train credit_bureau_b_1 ...\")\nt0 = time.time()\ntrain_features = build_cbb1_features(TRAIN_CBB1_PATHS)\nprint(f\"耗时: {time.time()-t0:.1f} 秒 | 内存: {train_features.estimated_size('mb'):.1f} MB\")\n\ntrain_features.write_parquet(OUT_TRAIN_PATH, compression=\"zstd\", compression_level=3)\ntrain_cols_list = train_features.columns\ndel train_features\ngc.collect()\n\n# ============================================================\n# Cell 5: 运行 Test\n# ============================================================\nprint(\"\\n>>> 开始处理 Test credit_bureau_b_1 ...\")\ntest_features = build_cbb1_features(TEST_CBB1_PATHS)\ntest_features.write_parquet(OUT_TEST_PATH, compression=\"zstd\", compression_level=3)\ntest_cols_list = test_features.columns\n\nmissing_in_test = set(train_cols_list) - set(test_cols_list)\nextra_in_test = set(test_cols_list) - set(train_cols_list)\n\ndel test_features\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:12.898073Z","iopub.execute_input":"2026-08-23T10:28:12.898353Z","iopub.status.idle":"2026-08-23T10:28:13.443564Z","shell.execute_reply.started":"2026-08-23T10:28:12.898327Z","shell.execute_reply":"2026-08-23T10:28:13.442711Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6: 写日志与清理缓存\n# ============================================================\nlog_content = f\"\"\"============================================================\n数据清洗日志: credit_bureau_b_1\n处理时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n\n[1. 字段丢弃]\n主动丢弃的无效字段 (0 个):\n无 (所有特征缺失率均未触发死刑红线)\n\n[2. 训练集与测试集形态]\nTrain 特征列数 : {len(train_cols_list)} 列\nTest  特征列数 : {len(test_cols_list)} 列\n\n[3. Schema 差异]\nTrain 有但 Test 没有的列 ({len(missing_in_test)} 个):\n{', '.join(list(missing_in_test)) if missing_in_test else '无'}\n\nTest 有但 Train 没有的列 ({len(extra_in_test)} 个):\n{', '.join(list(extra_in_test)) if extra_in_test else '无'}\n============================================================\n\"\"\"\nprint(log_content)\n\nlog_path = OUT_ROOT / \"credit_bureau_b_1_clean_log.txt\"\nwith open(log_path, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\n    \ngc.collect()\nprint(f\"🎉 credit_bureau_b_1 轻松搞定！日志已保存至: {log_path}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:13.444803Z","iopub.execute_input":"2026-08-23T10:28:13.445102Z","iopub.status.idle":"2026-08-23T10:28:13.526831Z","shell.execute_reply.started":"2026-08-23T10:28:13.445075Z","shell.execute_reply":"2026-08-23T10:28:13.525702Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# credit_bureau_b_2 表（征信局 B 的历史底层还款流水表）\n\ncredit_bureau_b_2（征信局 B 的历史底层还款流水表）的数据形态非常完美：\n\n体量小：只有 128 万行，不到 2MB，随便跑无压力。\n\n纯净度极高：一共只有 3 个业务字段（1 个日期，1 个逾期天数，1 个逾期金额），缺失率不到 0.5%，完全没有废字段。\n\n风控价值大：虽然小，但底层流水中的 dpd（逾期天数）和 overdue（逾期金额）依然是非常核心的信用特征。\n\n针对这种“小而美”的纯数值/日期流水表，我们直接全量加载，平行拍平（Depth 2 $\\rightarrow$ Depth 0），提取极值、总和以及逾期次数即可。\n","metadata":{}},{"cell_type":"code","source":"# ============================================================\n# Cell 1: 路径与环境准备 (credit_bureau_b_2)\n# ============================================================\nimport polars as pl\nimport gc\nimport time\nfrom pathlib import Path\nfrom datetime import datetime\n\n# 1. 定义数据源根目录 (Kaggle 环境)\nDATA_ROOT = Path(\"/kaggle/input/competitions/home-credit-credit-risk-model-stability/parquet_files\")\nTRAIN_DIR = DATA_ROOT / \"train\"\nTEST_DIR = DATA_ROOT / \"test\"\n\n# 2. 获取所有分片路径\nTRAIN_CBB2_PATHS = sorted(list(TRAIN_DIR.glob(\"train_credit_bureau_b_2*.parquet\")))\nTEST_CBB2_PATHS = sorted(list(TEST_DIR.glob(\"test_credit_bureau_b_2*.parquet\")))\n\n# 3. 定义输出目录\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nOUT_ROOT.mkdir(parents=True, exist_ok=True)\n\n# 4. 定义聚合后文件的保存路径\nOUT_TRAIN_PATH = OUT_ROOT / \"credit_bureau_b_2_clean_train.parquet\"\nOUT_TEST_PATH = OUT_ROOT / \"credit_bureau_b_2_clean_test.parquet\"\n\nprint(f\"✅ 环境准备完毕:\")\nprint(f\"   Train 分片数: {len(TRAIN_CBB2_PATHS)}\")\nprint(f\"   Test  分片数: {len(TEST_CBB2_PATHS)}\")\nassert len(TRAIN_CBB2_PATHS) > 0, \"❌ 未找到 train_credit_bureau_b_2\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:13.528300Z","iopub.execute_input":"2026-08-23T10:28:13.529147Z","iopub.status.idle":"2026-08-23T10:28:13.556615Z","shell.execute_reply.started":"2026-08-23T10:28:13.529111Z","shell.execute_reply":"2026-08-23T10:28:13.555618Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 2: 字段分类配置 (基于 credit_bureau_b_2 探查)\n# ============================================================\n\n# --- 主键 ---\nKEY_COLS = [\"case_id\", \"num_group1\", \"num_group2\"]\n\n# --- 绝对垃圾字段 ---\nDROP_COLS = []\n\n# --- 日期特征 (求 max/min) ---\nDATE_COLS = [\n    \"pmts_date_1107D\"\n]\n\n# --- 特殊风控字段：逾期天数 (提取严重逾期次数) ---\nDPD_COLS = [\n    \"pmts_dpdvalue_108P\"\n]\n\n# (将自动推断剩余数值特征，本表无类别特征)\nprint(f\"✅ 字段分类配置完成，将提取独立日期边界、金额及逾期统计值。\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:13.558088Z","iopub.execute_input":"2026-08-23T10:28:13.558408Z","iopub.status.idle":"2026-08-23T10:28:13.569778Z","shell.execute_reply.started":"2026-08-23T10:28:13.558370Z","shell.execute_reply":"2026-08-23T10:28:13.568176Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 3: 核心单表聚合函数 (极简版)\n# ============================================================\n\ndef build_cbb2_features(paths: list, verbose: bool = True) -> pl.DataFrame:\n    if verbose:\n        print(f\"[Step 1] 读取数据并采用宽松合并...\")\n    df = pl.concat([pl.scan_parquet(p) for p in paths], how=\"vertical_relaxed\")\n    \n    if verbose:\n        print(f\"[Step 2] 转换基础类型 (主键与日期)...\")\n    \n    df = df.with_columns(pl.col(\"case_id\").cast(pl.UInt32))\n    \n    for c in DATE_COLS:\n        df = df.with_columns(pl.col(c).str.to_date(\"%Y-%m-%d\", strict=False))\n        \n    # 动态获取数值列\n    current_cols = set(df.columns)\n    NUMERIC_COLS = list(current_cols - set(KEY_COLS + DATE_COLS + DROP_COLS))\n    \n    if verbose:\n        print(f\"[Step 3] 配置全局聚合算子...\")\n        \n    agg_exprs = [pl.col(\"case_id\").count().cast(pl.UInt16).alias(\"cbb2_total_payments\")]\n    \n    # 1. 日期列\n    for c in DATE_COLS:\n        agg_exprs.extend([\n            pl.col(c).min().alias(f\"{c}_min\"),\n            pl.col(c).max().alias(f\"{c}_max\"),\n        ])\n        \n    # 2. 数值列 (逾期金额和天数)\n    for c in NUMERIC_COLS:\n        agg_exprs.extend([\n            pl.col(c).max().alias(f\"{c}_max\"),\n            pl.col(c).mean().alias(f\"{c}_mean\"),\n            pl.col(c).sum().alias(f\"{c}_sum\"),\n            pl.col(c).is_not_null().sum().cast(pl.UInt16).alias(f\"{c}_notnull_cnt\"),\n        ])\n        \n    # 3. 逾期次数特别提取\n    for c in DPD_COLS:\n        agg_exprs.append(\n            (pl.col(c) > 0).sum().cast(pl.UInt16).alias(f\"{c}_gt0_cnt\")\n        )\n        \n    if verbose:\n        print(f\"[Step 4] 执行聚合...\")\n        \n    # 数据量小且没有透视，直接 group_by 即可\n    result = df.group_by(\"case_id\").agg(agg_exprs).collect(engine=\"streaming\")\n    \n    if verbose:\n        print(f\"✅ credit_bureau_b_2 清洗完毕，输出维度: {result.shape}\")\n        \n    return result","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:13.571890Z","iopub.execute_input":"2026-08-23T10:28:13.572208Z","iopub.status.idle":"2026-08-23T10:28:13.592237Z","shell.execute_reply.started":"2026-08-23T10:28:13.572180Z","shell.execute_reply":"2026-08-23T10:28:13.590879Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 4: 运行 Train 全量 (128 万行，秒级完成)\n# ============================================================\nprint(\">>> 开始处理 Train credit_bureau_b_2 ...\")\nt0 = time.time()\ntrain_features = build_cbb2_features(TRAIN_CBB2_PATHS)\nprint(f\"耗时: {time.time()-t0:.1f} 秒 | 内存: {train_features.estimated_size('mb'):.1f} MB\")\n\ntrain_features.write_parquet(OUT_TRAIN_PATH, compression=\"zstd\", compression_level=3)\ntrain_cols_list = train_features.columns\ndel train_features\ngc.collect()\n\n# ============================================================\n# Cell 5: 运行 Test\n# ============================================================\nprint(\"\\n>>> 开始处理 Test credit_bureau_b_2 ...\")\ntest_features = build_cbb2_features(TEST_CBB2_PATHS)\ntest_features.write_parquet(OUT_TEST_PATH, compression=\"zstd\", compression_level=3)\ntest_cols_list = test_features.columns\n\nmissing_in_test = set(train_cols_list) - set(test_cols_list)\nextra_in_test = set(test_cols_list) - set(train_cols_list)\n\ndel test_features\ngc.collect()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:13.593640Z","iopub.execute_input":"2026-08-23T10:28:13.594033Z","iopub.status.idle":"2026-08-23T10:28:13.886283Z","shell.execute_reply.started":"2026-08-23T10:28:13.593999Z","shell.execute_reply":"2026-08-23T10:28:13.885120Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# ============================================================\n# Cell 6: 写日志与清理缓存\n# ============================================================\nlog_content = f\"\"\"============================================================\n数据清洗日志: credit_bureau_b_2\n处理时间: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n============================================================\n\n[1. 字段丢弃]\n主动丢弃的无效字段 (0 个): 无\n\n[2. 训练集与测试集形态]\nTrain 特征列数 : {len(train_cols_list)} 列\nTest  特征列数 : {len(test_cols_list)} 列\n\n[3. Schema 差异]\nTrain 有但 Test 没有的列 ({len(missing_in_test)} 个):\n{', '.join(list(missing_in_test)) if missing_in_test else '无'}\n\nTest 有但 Train 没有的列 ({len(extra_in_test)} 个):\n{', '.join(list(extra_in_test)) if extra_in_test else '无'}\n============================================================\n\"\"\"\nprint(log_content)\n\nlog_path = OUT_ROOT / \"credit_bureau_b_2_clean_log.txt\"\nwith open(log_path, \"w\", encoding=\"utf-8\") as f:\n    f.write(log_content)\n    \ngc.collect()\nprint(f\"🎉 credit_bureau_b_2 处理完美收官！日志已保存至: {log_path}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:13.887634Z","iopub.execute_input":"2026-08-23T10:28:13.888042Z","iopub.status.idle":"2026-08-23T10:28:13.966979Z","shell.execute_reply.started":"2026-08-23T10:28:13.888001Z","shell.execute_reply":"2026-08-23T10:28:13.965691Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"from pathlib import Path\nOUT_ROOT = Path(\"/kaggle/working/clean\")\nprint(\"当前所有日志文件:\")\nfor f in sorted(OUT_ROOT.glob(\"*_log.txt\")):\n    size_kb = f.stat().st_size / 1024\n    print(f\"  {f.name}  ({size_kb:.1f} KB)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:33.159968Z","iopub.execute_input":"2026-08-23T10:28:33.164435Z","iopub.status.idle":"2026-08-23T10:28:33.194625Z","shell.execute_reply.started":"2026-08-23T10:28:33.163902Z","shell.execute_reply":"2026-08-23T10:28:33.190368Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"print(\"\\n所有 parquet 文件:\")\nfor f in sorted(OUT_ROOT.glob(\"*.parquet\")):\n    size_mb = f.stat().st_size / 1024**2\n    print(f\"  {f.name}  ({size_mb:.2f} MB)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:28:35.422283Z","iopub.execute_input":"2026-08-23T10:28:35.422878Z","iopub.status.idle":"2026-08-23T10:28:35.442913Z","shell.execute_reply.started":"2026-08-23T10:28:35.422804Z","shell.execute_reply":"2026-08-23T10:28:35.439239Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"from pathlib import Path\nOUT_ROOT = Path(\"/kaggle/working/clean\")\n\n# 重命名 applprev_1 让它和其他表统一\n(OUT_ROOT / \"train_applprev_1_agg.parquet\").rename(OUT_ROOT / \"applprev_1_clean_train.parquet\")\n(OUT_ROOT / \"test_applprev_1_agg.parquet\").rename(OUT_ROOT / \"applprev_1_clean_test.parquet\")\n\nprint(\"✅ 重命名完成,当前 applprev_1 相关文件:\")\nfor f in sorted(OUT_ROOT.glob(\"*applprev_1*\")):\n    print(f\"  {f.name}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-08-23T10:29:57.883711Z","iopub.execute_input":"2026-08-23T10:29:57.884161Z","iopub.status.idle":"2026-08-23T10:29:57.893436Z","shell.execute_reply.started":"2026-08-23T10:29:57.884125Z","shell.execute_reply":"2026-08-23T10:29:57.892444Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# base表：主动加字段（朴素表的特征加工）\n\n（1）base表的背景：\n\n整个比赛数据集的核心骨架。无论是 train_base 还是 test_base，它们的主要作用是提供每一笔信贷申请的最基础的时间和标识信息 ★\n\n主表(driving table)——所有其他 depth≥1 的表(applprev、person、credit_bureau...)清洗后最终都要 LEFT JOIN 到 base 上，所以真正需要做的不是\"压缩 base\",而是：\n\n\n> 1. 保留 base 的每一行不变(它已经是\"一 case_id 一行\")\n> 2. 在 base 上派生新字段,为后续 JOIN 和建模服务\n> 3. 对 date_decision 做时间特征展开,把日期信息榨干\n\n# --------------------------------------------\n\n（2）字段-1：case_id — 保留,不动\n\n(申请编号 / 案件 ID)：每一笔信贷申请的唯一标识符，也是整个数据集的主键。你需要用它来与所有其他带有历史记录的深度表（Depth=1 或 Depth=2）进行 join 关联\n\n理由:主键,后续所有 JOIN 都靠它\n\n注意事项:虽然唯一率 97.2%,但保险起见,建模前跑一次 n_unique() == height 的断言,防止意外重复\n# --------------------------------------------\n\n（3）字段 2:date_decision — 重点改造 ★★★ \n\n决策日期：向客户做出最终审批决策（如放款或拒绝）的具体日期；是极其关键的时间基准点；深度表中的各种历史时间字段（如过去某次逾期发生在多久前）都需要参考这个日期来计算时间差（天数/月数）\n\n当前：String 类型,596 个唯一值(约 596 天)\n\n问题：字符串日期对模型完全无用,必须转成数值特征。\n\n派生方案：后期朴素表的特征加工 ⭐️⭐️⭐️\n\n![屏幕截图 2026-08-21 171705.png](attachment:ec38fbb5-620d-47b2-b4d1-b6715a62833b.png)\n\n1.跑完 baseline 后，回头看2件事：\n\n> 根据 特征的重要性排序 feature_importance：排名靠后的直接删\n> \n> PSI指标标记 (train vs test)：> 0.25 的直接删\n>\n> decision_day：建模后如果 LightGBM feature_importance 里它排在末位,直接删掉；但 Home Credit 是消费贷放贷申请场景,申请日期在月内哪一天,和违约的关系没那么直接；；真实情况更可能是：客户在需要钱的那天去申请,这个\"需要钱的时间点\"是随机的；月末申请多可能只是因为工资到账后可以配首付,不代表风险更高\n> \n\n2.派生原理：\n\n特征工程\"和\"防止冗余/过拟合\"之间的天然矛盾——大多数教程都倾向多给,因为\"多给不会错\";但在稳定性竞赛里,少而精比多而杂更安全\n\n\n\n\n","metadata":{},"attachments":{"ec38fbb5-620d-47b2-b4d1-b6715a62833b.png":{"image/png":"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"}}},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}