{
  "id": 173849,
  "title": "Patient IDs in Test.csv and Train.csv",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/173849",
  "author_name": "",
  "post_date": "2020-08-11T01:59:11.802565700Z",
  "votes": null,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Please I need help to understand the Patient IDs.</p>\n<p>In the provided Train.csv and Test.csv, I can see that all patients in Test.csv also have records in Train.csv for multiple weeks.</p>\n<p><code>train[train.Patient.isin(test.Patient.unique())].shape</code><br>\ngives<br>\n(45, 7)</p>\n<p>But I think this is not true for the hidden data (atleast some patient id which is not in train.csv but present in test.csv hidden set). Is this assumption true or mentioned anywhere in the competition.</p>\n<p>I apologize if I have missed something here.</p>",
  "messages": [
    {
      "id": "965939",
      "postDate": "08/11/2020 01:59:11",
      "content": "<p>Please I need help to understand the Patient IDs.</p>\n<p>In the provided Train.csv and Test.csv, I can see that all patients in Test.csv also have records in Train.csv for multiple weeks.</p>\n<p><code>train[train.Patient.isin(test.Patient.unique())].shape</code><br>\ngives<br>\n(45, 7)</p>\n<p>But I think this is not true for the hidden data (atleast some patient id which is not in train.csv but present in test.csv hidden set). Is this assumption true or mentioned anywhere in the competition.</p>\n<p>I apologize if I have missed something here.</p>",
      "rawMarkdown": "Please I need help to understand the Patient IDs.\n\nIn the provided Train.csv and Test.csv, I can see that all patients in Test.csv also have records in Train.csv for multiple weeks.\n\n`train[train.Patient.isin(test.Patient.unique())].shape`\ngives\n(45, 7)\n\nBut I think this is not true for the hidden data (atleast some patient id which is not in train.csv but present in test.csv hidden set). Is this assumption true or mentioned anywhere in the competition.\n\nI apologize if I have missed something here.",
      "votes": null
    },
    {
      "id": "965945",
      "postDate": "08/11/2020 02:03:24",
      "content": "<p>The competition creators have mentioned that the Patient IDs in test are just placeholders. So there should be no expectation that private will have overlapping Patient IDs between train and test.</p>",
      "rawMarkdown": "The competition creators have mentioned that the Patient IDs in test are just placeholders. So there should be no expectation that private will have overlapping Patient IDs between train and test.",
      "votes": null
    },
    {
      "id": "965958",
      "postDate": "08/11/2020 02:37:58",
      "content": "<p>Thanks a lot for this quick clarification.</p>",
      "rawMarkdown": "Thanks a lot for this quick clarification.",
      "votes": null
    },
    {
      "id": "966647",
      "postDate": "08/11/2020 15:37:14",
      "content": "<p>The test csv is only a guide for you to shape your result in a way that would be compatible for the hidden test set. So, it does not matter if patient ids from train are in the test, since anyway, those patient ids will not be used for actual testing.  </p>",
      "rawMarkdown": "The test csv is only a guide for you to shape your result in a way that would be compatible for the hidden test set. So, it does not matter if patient ids from train are in the test, since anyway, those patient ids will not be used for actual testing.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 965945,
      "author_name": "jjinho",
      "author_url": "",
      "post_date": "08/11/2020 02:03:24",
      "content": "<p>The competition creators have mentioned that the Patient IDs in test are just placeholders. So there should be no expectation that private will have overlapping Patient IDs between train and test.</p>",
      "votes": null,
      "replies": [
        {
          "id": 965958,
          "author_name": "aliasgherman",
          "author_url": "",
          "post_date": "08/11/2020 02:37:58",
          "content": "<p>Thanks a lot for this quick clarification.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 966647,
      "author_name": "kitangladmatu",
      "author_url": "",
      "post_date": "08/11/2020 15:37:14",
      "content": "<p>The test csv is only a guide for you to shape your result in a way that would be compatible for the hidden test set. So, it does not matter if patient ids from train are in the test, since anyway, those patient ids will not be used for actual testing.  </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "965939": "Please I need help to understand the Patient IDs.\n\nIn the provided Train.csv and Test.csv, I can see that all patients in Test.csv also have records in Train.csv for multiple weeks.\n\n`train[train.Patient.isin(test.Patient.unique())].shape`\ngives\n(45, 7)\n\nBut I think this is not true for the hidden data (atleast some patient id which is not in train.csv but present in test.csv hidden set). Is this assumption true or mentioned anywhere in the competition.\n\nI apologize if I have missed something here.",
    "965945": "The competition creators have mentioned that the Patient IDs in test are just placeholders. So there should be no expectation that private will have overlapping Patient IDs between train and test.",
    "965958": "Thanks a lot for this quick clarification.",
    "966647": "The test csv is only a guide for you to shape your result in a way that would be compatible for the hidden test set. So, it does not matter if patient ids from train are in the test, since anyway, those patient ids will not be used for actual testing."
  },
  "source": "meta"
}