{
  "id": 193030,
  "title": "Score Error Cause",
  "url": "/competitions/riiid-test-answer-prediction/discussion/193030",
  "author_name": "",
  "post_date": "2020-10-24T23:15:51.513178400Z",
  "votes": 5,
  "comment_count": 2,
  "views": 0,
  "content": "<p>If you are still having this problem, this is a possible fix I found out in my case.<br>\nWhen doing pd.Merge, my code didnt specify the 'how' method. This caused Pandas to mix up the Row Id index in the submission.csv. I noticed it after a commit. The fix to this was using <br>\npd.merge(… , how='left'), and I was able to submit.</p>\n<p>So if you were able to commit and got this error, check out the submission.csv file and see if the Row ID is correctly in order from 0 and on. Even if you have all predictions, you will still get the error if they are out of order. </p>",
  "messages": [
    {
      "id": "1059319",
      "postDate": "10/24/2020 23:15:51",
      "content": "<p>If you are still having this problem, this is a possible fix I found out in my case.<br>\nWhen doing pd.Merge, my code didnt specify the 'how' method. This caused Pandas to mix up the Row Id index in the submission.csv. I noticed it after a commit. The fix to this was using <br>\npd.merge(… , how='left'), and I was able to submit.</p>\n<p>So if you were able to commit and got this error, check out the submission.csv file and see if the Row ID is correctly in order from 0 and on. Even if you have all predictions, you will still get the error if they are out of order. </p>",
      "rawMarkdown": "If you are still having this problem, this is a possible fix I found out in my case.\nWhen doing pd.Merge, my code didnt specify the 'how' method. This caused Pandas to mix up the Row Id index in the submission.csv. I noticed it after a commit. The fix to this was using \npd.merge(... , how='left'), and I was able to submit.\n\nSo if you were able to commit and got this error, check out the submission.csv file and see if the Row ID is correctly in order from 0 and on. Even if you have all predictions, you will still get the error if they are out of order.",
      "votes": null
    },
    {
      "id": "1059524",
      "postDate": "10/25/2020 07:48:18",
      "content": "<p>Thanks, Luryck, glad you solved that mystery. My annoying Score Errors were due to invisible trailing tab characters in read_csv input files. They caused numerical values to be parsed as NaN. \"Run All\" in the Kaggle Notebook window reported no problems, but Submissions failed.</p>",
      "rawMarkdown": "Thanks, Luryck, glad you solved that mystery. My annoying Score Errors were due to invisible trailing tab characters in read_csv input files. They caused numerical values to be parsed as NaN. \"Run All\" in the Kaggle Notebook window reported no problems, but Submissions failed.",
      "votes": null
    },
    {
      "id": "1059973",
      "postDate": "10/25/2020 16:29:17",
      "content": "<p>I can only imagine the pain to debug that</p>",
      "rawMarkdown": "I can only imagine the pain to debug that",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1059524,
      "author_name": "mikel1",
      "author_url": "",
      "post_date": "10/25/2020 07:48:18",
      "content": "<p>Thanks, Luryck, glad you solved that mystery. My annoying Score Errors were due to invisible trailing tab characters in read_csv input files. They caused numerical values to be parsed as NaN. \"Run All\" in the Kaggle Notebook window reported no problems, but Submissions failed.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1059973,
          "author_name": "iuryck",
          "author_url": "",
          "post_date": "10/25/2020 16:29:17",
          "content": "<p>I can only imagine the pain to debug that</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1059319": "If you are still having this problem, this is a possible fix I found out in my case.\nWhen doing pd.Merge, my code didnt specify the 'how' method. This caused Pandas to mix up the Row Id index in the submission.csv. I noticed it after a commit. The fix to this was using \npd.merge(... , how='left'), and I was able to submit.\n\nSo if you were able to commit and got this error, check out the submission.csv file and see if the Row ID is correctly in order from 0 and on. Even if you have all predictions, you will still get the error if they are out of order.",
    "1059524": "Thanks, Luryck, glad you solved that mystery. My annoying Score Errors were due to invisible trailing tab characters in read_csv input files. They caused numerical values to be parsed as NaN. \"Run All\" in the Kaggle Notebook window reported no problems, but Submissions failed.",
    "1059973": "I can only imagine the pain to debug that"
  },
  "source": "meta"
}