{
  "id": 58405,
  "title": "event_id error in submission",
  "url": "/competitions/trackml-particle-identification/discussion/58405",
  "author_name": "",
  "post_date": "2018-06-07T13:31:07.494871900Z",
  "votes": 2,
  "comment_count": 9,
  "views": 0,
  "content": "<p>I'm facing a technical issue when submitting my csv file to score. This warning shows up at the very end of the scoring process:\n<code>Evaluation Exception: Submission does not contain correct event_ids. See sample_submission for correct values.</code>\nI downloaded both sample_submission and my submission.csv to my local machine, just to check everything locally. All event_ids are correct, ranging from 0 to 124, and the number of rows is correct as well. \n<img src=\"http://image.ibb.co/mj00Xo/submission.png\" alt=\"sample and actual submission comparison\">\nDoes anyone know what may be causing this?</p>\n\n<p>EDIT: Just in case anyone find themselves facing this issue, it was solved by sorting the submission file by event_id. Mine was beginning at 124 going down to 0. It seems the score algorithm doesn't like that :P So just\n<code>submission.sort_values(by = [\"event_id\", \"hit_id\"], ascending=True).reset_index()</code></p>",
  "messages": [
    {
      "id": "339716",
      "postDate": "06/07/2018 13:31:07",
      "content": "<p>I'm facing a technical issue when submitting my csv file to score. This warning shows up at the very end of the scoring process:\n<code>Evaluation Exception: Submission does not contain correct event_ids. See sample_submission for correct values.</code>\nI downloaded both sample_submission and my submission.csv to my local machine, just to check everything locally. All event_ids are correct, ranging from 0 to 124, and the number of rows is correct as well. \n<img src=\"http://image.ibb.co/mj00Xo/submission.png\" alt=\"sample and actual submission comparison\">\nDoes anyone know what may be causing this?</p>\n\n<p>EDIT: Just in case anyone find themselves facing this issue, it was solved by sorting the submission file by event_id. Mine was beginning at 124 going down to 0. It seems the score algorithm doesn't like that :P So just\n<code>submission.sort_values(by = [\"event_id\", \"hit_id\"], ascending=True).reset_index()</code></p>",
      "rawMarkdown": "I'm facing a technical issue when submitting my csv file to score. This warning shows up at the very end of the scoring process:\n`Evaluation Exception: Submission does not contain correct event_ids. See sample_submission for correct values.`\nI downloaded both sample_submission and my submission.csv to my local machine, just to check everything locally. All event_ids are correct, ranging from 0 to 124, and the number of rows is correct as well. \n![sample and actual submission comparison][1]\n[1]: http://image.ibb.co/mj00Xo/submission.png\nDoes anyone know what may be causing this?\n\nEDIT: Just in case anyone find themselves facing this issue, it was solved by sorting the submission file by event_id. Mine was beginning at 124 going down to 0. It seems the score algorithm doesn't like that :P So just\n`submission.sort_values(by = [\"event_id\", \"hit_id\"], ascending=True).reset_index()`",
      "votes": null
    },
    {
      "id": "339734",
      "postDate": "06/07/2018 14:27:55",
      "content": "<p>You may have a corrupted upload.  Try submitting the same content with a different filename to be sure.  I recommend you compress your submission, 7z is the one yielding smallest file sizes.</p>\n\n<p>Edit: you should also check that the number of occurrence of each event value is the same in your sub and in the sample sub.</p>",
      "rawMarkdown": "You may have a corrupted upload.  Try submitting the same content with a different filename to be sure.  I recommend you compress your submission, 7z is the one yielding smallest file sizes.\n\nEdit: you should also check that the number of occurrence of each event value is the same in your sub and in the sample sub.",
      "votes": null
    },
    {
      "id": "339783",
      "postDate": "06/07/2018 16:27:53",
      "content": "<p>Thanks for your response! Actually, I submitted my file directly from the Kaggle kernel, so it shouldn't be a file corruption issue. I also sliced the submission file for each event and compared its size with sample_submission. The number of occurrence in each one (sample and mine) are the same for all events. Don't know what to do anymore :(</p>",
      "rawMarkdown": "Thanks for your response! Actually, I submitted my file directly from the Kaggle kernel, so it shouldn't be a file corruption issue. I also sliced the submission file for each event and compared its size with sample_submission. The number of occurrence in each one (sample and mine) are the same for all events. Don't know what to do anymore :(",
      "votes": null
    },
    {
      "id": "339790",
      "postDate": "06/07/2018 17:01:29",
      "content": "<blockquote>\n  <p>so it shouldn't be a file corruption issue</p>\n</blockquote>\n\n<p>I'd check it still.  If everything always worked as it should in IT then life would be way different.</p>",
      "rawMarkdown": "&gt;  so it shouldn't be a file corruption issue\n\nI'd check it still.  If everything always worked as it should in IT then life would be way different.",
      "votes": null
    },
    {
      "id": "339794",
      "postDate": "06/07/2018 17:10:32",
      "content": "<p>hahaha true. I'm running the kernel again. Maybe this time it'll work.</p>\n\n<p>EDIT: Ran the kernel again and same error :/</p>\n\n<p>EDIT 2: Did it! I wrote how I fixed it on the main post. Thanks for your support, CPMP!</p>",
      "rawMarkdown": "hahaha true. I'm running the kernel again. Maybe this time it'll work.\n\nEDIT: Ran the kernel again and same error :/\n\nEDIT 2: Did it! I wrote how I fixed it on the main post. Thanks for your support, CPMP!",
      "votes": null
    },
    {
      "id": "341599",
      "postDate": "06/11/2018 21:27:23",
      "content": "<p>Thanks for following up. Added one line about it in Welcome from organizers</p>",
      "rawMarkdown": "Thanks for following up. Added one line about it in Welcome from organizers",
      "votes": null
    },
    {
      "id": "341674",
      "postDate": "06/12/2018 04:28:20",
      "content": "<p>You should add this information in the submission page or in the description of the sample data in the data page too.</p>",
      "rawMarkdown": "You should add this information in the submission page or in the description of the sample data in the data page too.",
      "votes": null
    },
    {
      "id": "341715",
      "postDate": "06/12/2018 05:44:28",
      "content": "<p>Nice catch, really wonder why the scoring function depends on the order of rows in the submission file.  First time I am seeing this at Kaggle.</p>",
      "rawMarkdown": "Nice catch, really wonder why the scoring function depends on the order of rows in the submission file.  First time I am seeing this at Kaggle.",
      "votes": null
    },
    {
      "id": "341784",
      "postDate": "06/12/2018 09:09:01",
      "content": "<p>Happened in \"Flavours of Physics: Finding τ → μμμ\"</p>",
      "rawMarkdown": "Happened in \"Flavours of Physics: Finding τ → μμμ\"",
      "votes": null
    },
    {
      "id": "341920",
      "postDate": "06/12/2018 14:25:08",
      "content": "<blockquote>\n  <p>Happened in \"Flavours of Physics: Finding τ → μμμ\"</p>\n</blockquote>\n\n<p>That's before I started Kaggle competitions ;)</p>\n\n<p>Interestingly it comes from the same organizers.  We can therefore infer: CERN competitions evaluation are dependent on the row order in submission files ;)</p>",
      "rawMarkdown": "&gt; Happened in \"Flavours of Physics: Finding τ → μμμ\"\n\nThat's before I started Kaggle competitions ;)\n\nInterestingly it comes from the same organizers.  We can therefore infer: CERN competitions evaluation are dependent on the row order in submission files ;)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 339734,
      "author_name": "cpmpml",
      "author_url": "",
      "post_date": "06/07/2018 14:27:55",
      "content": "<p>You may have a corrupted upload.  Try submitting the same content with a different filename to be sure.  I recommend you compress your submission, 7z is the one yielding smallest file sizes.</p>\n\n<p>Edit: you should also check that the number of occurrence of each event value is the same in your sub and in the sample sub.</p>",
      "votes": null,
      "replies": [
        {
          "id": 339783,
          "author_name": "hrmello",
          "author_url": "",
          "post_date": "06/07/2018 16:27:53",
          "content": "<p>Thanks for your response! Actually, I submitted my file directly from the Kaggle kernel, so it shouldn't be a file corruption issue. I also sliced the submission file for each event and compared its size with sample_submission. The number of occurrence in each one (sample and mine) are the same for all events. Don't know what to do anymore :(</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 339790,
          "author_name": "cpmpml",
          "author_url": "",
          "post_date": "06/07/2018 17:01:29",
          "content": "<blockquote>\n  <p>so it shouldn't be a file corruption issue</p>\n</blockquote>\n\n<p>I'd check it still.  If everything always worked as it should in IT then life would be way different.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 339794,
          "author_name": "hrmello",
          "author_url": "",
          "post_date": "06/07/2018 17:10:32",
          "content": "<p>hahaha true. I'm running the kernel again. Maybe this time it'll work.</p>\n\n<p>EDIT: Ran the kernel again and same error :/</p>\n\n<p>EDIT 2: Did it! I wrote how I fixed it on the main post. Thanks for your support, CPMP!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 341715,
          "author_name": "cpmpml",
          "author_url": "",
          "post_date": "06/12/2018 05:44:28",
          "content": "<p>Nice catch, really wonder why the scoring function depends on the order of rows in the submission file.  First time I am seeing this at Kaggle.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 341784,
          "author_name": "glimmung",
          "author_url": "",
          "post_date": "06/12/2018 09:09:01",
          "content": "<p>Happened in \"Flavours of Physics: Finding τ → μμμ\"</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 341920,
          "author_name": "cpmpml",
          "author_url": "",
          "post_date": "06/12/2018 14:25:08",
          "content": "<blockquote>\n  <p>Happened in \"Flavours of Physics: Finding τ → μμμ\"</p>\n</blockquote>\n\n<p>That's before I started Kaggle competitions ;)</p>\n\n<p>Interestingly it comes from the same organizers.  We can therefore infer: CERN competitions evaluation are dependent on the row order in submission files ;)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 341599,
      "author_name": "droussea",
      "author_url": "",
      "post_date": "06/11/2018 21:27:23",
      "content": "<p>Thanks for following up. Added one line about it in Welcome from organizers</p>",
      "votes": null,
      "replies": [
        {
          "id": 341674,
          "author_name": "cpmpml",
          "author_url": "",
          "post_date": "06/12/2018 04:28:20",
          "content": "<p>You should add this information in the submission page or in the description of the sample data in the data page too.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "339716": "I'm facing a technical issue when submitting my csv file to score. This warning shows up at the very end of the scoring process:\n`Evaluation Exception: Submission does not contain correct event_ids. See sample_submission for correct values.`\nI downloaded both sample_submission and my submission.csv to my local machine, just to check everything locally. All event_ids are correct, ranging from 0 to 124, and the number of rows is correct as well. \n![sample and actual submission comparison][1]\n[1]: http://image.ibb.co/mj00Xo/submission.png\nDoes anyone know what may be causing this?\n\nEDIT: Just in case anyone find themselves facing this issue, it was solved by sorting the submission file by event_id. Mine was beginning at 124 going down to 0. It seems the score algorithm doesn't like that :P So just\n`submission.sort_values(by = [\"event_id\", \"hit_id\"], ascending=True).reset_index()`",
    "339734": "You may have a corrupted upload.  Try submitting the same content with a different filename to be sure.  I recommend you compress your submission, 7z is the one yielding smallest file sizes.\n\nEdit: you should also check that the number of occurrence of each event value is the same in your sub and in the sample sub.",
    "339783": "Thanks for your response! Actually, I submitted my file directly from the Kaggle kernel, so it shouldn't be a file corruption issue. I also sliced the submission file for each event and compared its size with sample_submission. The number of occurrence in each one (sample and mine) are the same for all events. Don't know what to do anymore :(",
    "339790": "&gt;  so it shouldn't be a file corruption issue\n\nI'd check it still.  If everything always worked as it should in IT then life would be way different.",
    "339794": "hahaha true. I'm running the kernel again. Maybe this time it'll work.\n\nEDIT: Ran the kernel again and same error :/\n\nEDIT 2: Did it! I wrote how I fixed it on the main post. Thanks for your support, CPMP!",
    "341599": "Thanks for following up. Added one line about it in Welcome from organizers",
    "341674": "You should add this information in the submission page or in the description of the sample data in the data page too.",
    "341715": "Nice catch, really wonder why the scoring function depends on the order of rows in the submission file.  First time I am seeing this at Kaggle.",
    "341784": "Happened in \"Flavours of Physics: Finding τ → μμμ\"",
    "341920": "&gt; Happened in \"Flavours of Physics: Finding τ → μμμ\"\n\nThat's before I started Kaggle competitions ;)\n\nInterestingly it comes from the same organizers.  We can therefore infer: CERN competitions evaluation are dependent on the row order in submission files ;)"
  },
  "source": "meta"
}