{
  "id": 54531,
  "title": "test & supplement index mapping",
  "url": "/competitions/talkingdata-adtracking-fraud-detection/discussion/54531",
  "author_name": "",
  "post_date": "2018-04-14T13:31:17.431929Z",
  "votes": 6,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Here's my finding about test &amp; supplement overlapping samples index mapping summarized by the following figure.\nAnd some important numbers:</p>\n\n<ul>\n<li>offsets: [21290876, 29475763, 35793790]</li>\n<li>test head indexes: [0, 6202933, 12316147]</li>\n<li>test tail indexes: [6202932, 12316146, 18790468]</li>\n</ul>\n\n<p>This is simply based on some observations, so please tell me if there's something wrong~\nThe detailed contents are in the kernel: </p>\n\n<p><a href=\"https://www.kaggle.com/johnfarrell/tda-overlapping-sample-indexes-in-supplement\">https://www.kaggle.com/johnfarrell/tda-overlapping-sample-indexes-in-supplement</a></p>\n\n<p>Processed data download here:</p>\n\n<p><a href=\"https://www.kaggle.com/johnfarrell/talkingdata-adtracking-test-order-supplement\">https://www.kaggle.com/johnfarrell/talkingdata-adtracking-test-order-supplement</a></p>\n\n<p><img src=\"https://i.imgur.com/d0wC7Qm.png\" alt=\"img\"></p>",
  "messages": [
    {
      "id": "314041",
      "postDate": "04/14/2018 13:31:17",
      "content": "<p>Here's my finding about test &amp; supplement overlapping samples index mapping summarized by the following figure.\nAnd some important numbers:</p>\n\n<ul>\n<li>offsets: [21290876, 29475763, 35793790]</li>\n<li>test head indexes: [0, 6202933, 12316147]</li>\n<li>test tail indexes: [6202932, 12316146, 18790468]</li>\n</ul>\n\n<p>This is simply based on some observations, so please tell me if there's something wrong~\nThe detailed contents are in the kernel: </p>\n\n<p><a href=\"https://www.kaggle.com/johnfarrell/tda-overlapping-sample-indexes-in-supplement\">https://www.kaggle.com/johnfarrell/tda-overlapping-sample-indexes-in-supplement</a></p>\n\n<p>Processed data download here:</p>\n\n<p><a href=\"https://www.kaggle.com/johnfarrell/talkingdata-adtracking-test-order-supplement\">https://www.kaggle.com/johnfarrell/talkingdata-adtracking-test-order-supplement</a></p>\n\n<p><img src=\"https://i.imgur.com/d0wC7Qm.png\" alt=\"img\"></p>",
      "rawMarkdown": "Here's my finding about test &amp; supplement overlapping samples index mapping summarized by the following figure.\nAnd some important numbers:\n\n- offsets: [21290876, 29475763, 35793790]\n- test head indexes: [0, 6202933, 12316147]\n- test tail indexes: [6202932, 12316146, 18790468]\n\nThis is simply based on some observations, so please tell me if there's something wrong~\nThe detailed contents are in the kernel: \n\nhttps://www.kaggle.com/johnfarrell/tda-overlapping-sample-indexes-in-supplement\n\nProcessed data download here:\n\nhttps://www.kaggle.com/johnfarrell/talkingdata-adtracking-test-order-supplement\n\n![img][1]\n\n\n  [1]: https://i.imgur.com/d0wC7Qm.png",
      "votes": null
    },
    {
      "id": "314069",
      "postDate": "04/14/2018 14:43:40",
      "content": "<p>Yes, test is a subset of test_supplement, obtained by selecting some hours.</p>",
      "rawMarkdown": "Yes, test is a subset of test_supplement, obtained by selecting some hours.",
      "votes": null
    },
    {
      "id": "314081",
      "postDate": "04/14/2018 14:57:52",
      "content": "<p>I'm just curious about the direct index mapping in two csv files~ for the samples seem to be in slightly different order within the same second</p>",
      "rawMarkdown": "I'm just curious about the direct index mapping in two csv files~ for the samples seem to be in slightly different order within the same second",
      "votes": null
    },
    {
      "id": "314084",
      "postDate": "04/14/2018 15:01:56",
      "content": "<p>Yes, they are not in same order.  </p>",
      "rawMarkdown": "Yes, they are not in same order.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 314069,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "04/14/2018 14:43:40",
      "content": "<p>Yes, test is a subset of test_supplement, obtained by selecting some hours.</p>",
      "votes": null,
      "replies": [
        {
          "id": 314081,
          "author_name": "johnfarrell",
          "author_url": "",
          "post_date": "04/14/2018 14:57:52",
          "content": "<p>I'm just curious about the direct index mapping in two csv files~ for the samples seem to be in slightly different order within the same second</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 314084,
          "author_name": "cpmpml",
          "author_url": "",
          "post_date": "04/14/2018 15:01:56",
          "content": "<p>Yes, they are not in same order.  </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "314041": "Here's my finding about test &amp; supplement overlapping samples index mapping summarized by the following figure.\nAnd some important numbers:\n\n- offsets: [21290876, 29475763, 35793790]\n- test head indexes: [0, 6202933, 12316147]\n- test tail indexes: [6202932, 12316146, 18790468]\n\nThis is simply based on some observations, so please tell me if there's something wrong~\nThe detailed contents are in the kernel: \n\nhttps://www.kaggle.com/johnfarrell/tda-overlapping-sample-indexes-in-supplement\n\nProcessed data download here:\n\nhttps://www.kaggle.com/johnfarrell/talkingdata-adtracking-test-order-supplement\n\n![img][1]\n\n\n  [1]: https://i.imgur.com/d0wC7Qm.png",
    "314069": "Yes, test is a subset of test_supplement, obtained by selecting some hours.",
    "314081": "I'm just curious about the direct index mapping in two csv files~ for the samples seem to be in slightly different order within the same second",
    "314084": "Yes, they are not in same order."
  },
  "source": "meta"
}