{
  "id": 194406,
  "title": "Order of user interactions",
  "url": "/competitions/riiid-test-answer-prediction/discussion/194406",
  "author_name": "",
  "post_date": "2020-11-01T17:05:03.762446500Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>First off, sorry for rehashing old topics.</p>\n<p>I remember there was some discussion around whether user interaction timestamps in the test set follow directly after train set data. I can't seem to find it anymore. Can anyone help me reach that discussion? </p>\n<p>Also, was it mentioned somewhere that timestamps in the test set could be missing/NAvalues?</p>",
  "messages": [
    {
      "id": "1066366",
      "postDate": "11/01/2020 17:05:03",
      "content": "<p>First off, sorry for rehashing old topics.</p>\n<p>I remember there was some discussion around whether user interaction timestamps in the test set follow directly after train set data. I can't seem to find it anymore. Can anyone help me reach that discussion? </p>\n<p>Also, was it mentioned somewhere that timestamps in the test set could be missing/NAvalues?</p>",
      "rawMarkdown": "First off, sorry for rehashing old topics.\n\nI remember there was some discussion around whether user interaction timestamps in the test set follow directly after train set data. I can't seem to find it anymore. Can anyone help me reach that discussion? \n\nAlso, was it mentioned somewhere that timestamps in the test set could be missing/NAvalues?",
      "votes": null
    },
    {
      "id": "1066374",
      "postDate": "11/01/2020 17:17:19",
      "content": "<blockquote>\n  <p>I remember there was some discussion around whether user interaction timestamps in the test set follow directly after train set data. I can't seem to find it anymore. Can anyone help me reach that discussion?</p>\n</blockquote>\n<p>Fourth point <a href=\"https://www.kaggle.com/c/riiid-test-answer-prediction/discussion/191106\" target=\"_blank\">here</a>.</p>\n<blockquote>\n  <p>was it mentioned somewhere that timestamps in the test set could be missing/NAvalues?</p>\n</blockquote>\n<p>I don't think so but haven't tested on the private test set. If 101M rows of train data have no NAs, its very unlikely the private test set would.</p>",
      "rawMarkdown": "> I remember there was some discussion around whether user interaction timestamps in the test set follow directly after train set data. I can't seem to find it anymore. Can anyone help me reach that discussion?\n\nFourth point [here](https://www.kaggle.com/c/riiid-test-answer-prediction/discussion/191106).\n\n> was it mentioned somewhere that timestamps in the test set could be missing/NAvalues?\n\nI don't think so but haven't tested on the private test set. If 101M rows of train data have no NAs, its very unlikely the private test set would.",
      "votes": null
    },
    {
      "id": "1066375",
      "postDate": "11/01/2020 17:24:29",
      "content": "<p>Thanks  a lot <a href=\"https://www.kaggle.com/rohanrao\" target=\"_blank\">@rohanrao</a> :)</p>",
      "rawMarkdown": "Thanks  a lot @rohanrao :)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1066374,
      "author_name": "rohanrao",
      "author_url": "",
      "post_date": "11/01/2020 17:17:19",
      "content": "<blockquote>\n  <p>I remember there was some discussion around whether user interaction timestamps in the test set follow directly after train set data. I can't seem to find it anymore. Can anyone help me reach that discussion?</p>\n</blockquote>\n<p>Fourth point <a href=\"https://www.kaggle.com/c/riiid-test-answer-prediction/discussion/191106\" target=\"_blank\">here</a>.</p>\n<blockquote>\n  <p>was it mentioned somewhere that timestamps in the test set could be missing/NAvalues?</p>\n</blockquote>\n<p>I don't think so but haven't tested on the private test set. If 101M rows of train data have no NAs, its very unlikely the private test set would.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1066375,
          "author_name": "abhimanyud",
          "author_url": "",
          "post_date": "11/01/2020 17:24:29",
          "content": "<p>Thanks  a lot <a href=\"https://www.kaggle.com/rohanrao\" target=\"_blank\">@rohanrao</a> :)</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1066366": "First off, sorry for rehashing old topics.\n\nI remember there was some discussion around whether user interaction timestamps in the test set follow directly after train set data. I can't seem to find it anymore. Can anyone help me reach that discussion? \n\nAlso, was it mentioned somewhere that timestamps in the test set could be missing/NAvalues?",
    "1066374": "> I remember there was some discussion around whether user interaction timestamps in the test set follow directly after train set data. I can't seem to find it anymore. Can anyone help me reach that discussion?\n\nFourth point [here](https://www.kaggle.com/c/riiid-test-answer-prediction/discussion/191106).\n\n> was it mentioned somewhere that timestamps in the test set could be missing/NAvalues?\n\nI don't think so but haven't tested on the private test set. If 101M rows of train data have no NAs, its very unlikely the private test set would.",
    "1066375": "Thanks  a lot @rohanrao :)"
  },
  "source": "meta"
}