{
  "id": 328324,
  "title": "Identical Markdown cells with different ids in the same notebook",
  "url": "/competitions/AI4Code/discussion/328324",
  "author_name": "",
  "post_date": "2022-05-31T22:39:24.213916700Z",
  "votes": 10,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Hi,</p>\n<p>Doing the <a href=\"https://www.kaggle.com/code/josecarmona/ai4code-eda-and-notes-2\" target=\"_blank\">my competition EDA</a> I have found markdown cells with different id but with equal md5 hash, that means, the same markdown content. Examples in notebooks ids 807710d20e6842, edd4dbbd9e6d38, edd4dbbd9e6d38, etc.</p>\n<p>I think this is no problem in model training time or predicting, but I think this may be a problem in submission time because with Kendall Tau metric the order matters, obviously.</p>\n<p>Does anyone knows any rule in order to compose the submission file in perfect order for this case?</p>",
  "messages": [
    {
      "id": "1807300",
      "postDate": "05/31/2022 22:39:24",
      "content": "<p>Hi,</p>\n<p>Doing the <a href=\"https://www.kaggle.com/code/josecarmona/ai4code-eda-and-notes-2\" target=\"_blank\">my competition EDA</a> I have found markdown cells with different id but with equal md5 hash, that means, the same markdown content. Examples in notebooks ids 807710d20e6842, edd4dbbd9e6d38, edd4dbbd9e6d38, etc.</p>\n<p>I think this is no problem in model training time or predicting, but I think this may be a problem in submission time because with Kendall Tau metric the order matters, obviously.</p>\n<p>Does anyone knows any rule in order to compose the submission file in perfect order for this case?</p>",
      "rawMarkdown": "Hi,\n\nDoing the [my competition EDA](https://www.kaggle.com/code/josecarmona/ai4code-eda-and-notes-2) I have found markdown cells with different id but with equal md5 hash, that means, the same markdown content. Examples in notebooks ids 807710d20e6842, edd4dbbd9e6d38, edd4dbbd9e6d38, etc.\n\nI think this is no problem in model training time or predicting, but I think this may be a problem in submission time because with Kendall Tau metric the order matters, obviously.\n\nDoes anyone knows any rule in order to compose the submission file in perfect order for this case?",
      "votes": null
    },
    {
      "id": "1809100",
      "postDate": "06/02/2022 11:47:04",
      "content": "<p>Nice spot! If the content of the cells is identical, then I guess we're stuck, and the relative order of these particular subsets of cells will just have to be written off as noise. </p>",
      "rawMarkdown": "Nice spot! If the content of the cells is identical, then I guess we're stuck, and the relative order of these particular subsets of cells will just have to be written off as noise.",
      "votes": null
    },
    {
      "id": "1813752",
      "postDate": "06/07/2022 07:27:29",
      "content": "<p>In real life, this can happen. I mean, I can put headlines like \"EDA\" as markdowns in different parts of the notebook. Whether we could mark them as the noise completely depends on the frequency of the cells in the dataset.</p>",
      "rawMarkdown": "In real life, this can happen. I mean, I can put headlines like \"EDA\" as markdowns in different parts of the notebook. Whether we could mark them as the noise completely depends on the frequency of the cells in the dataset.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1809100,
      "author_name": "jagofc",
      "author_url": "",
      "post_date": "06/02/2022 11:47:04",
      "content": "<p>Nice spot! If the content of the cells is identical, then I guess we're stuck, and the relative order of these particular subsets of cells will just have to be written off as noise. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1813752,
      "author_name": "ashrafr",
      "author_url": "",
      "post_date": "06/07/2022 07:27:29",
      "content": "<p>In real life, this can happen. I mean, I can put headlines like \"EDA\" as markdowns in different parts of the notebook. Whether we could mark them as the noise completely depends on the frequency of the cells in the dataset.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1807300": "Hi,\n\nDoing the [my competition EDA](https://www.kaggle.com/code/josecarmona/ai4code-eda-and-notes-2) I have found markdown cells with different id but with equal md5 hash, that means, the same markdown content. Examples in notebooks ids 807710d20e6842, edd4dbbd9e6d38, edd4dbbd9e6d38, etc.\n\nI think this is no problem in model training time or predicting, but I think this may be a problem in submission time because with Kendall Tau metric the order matters, obviously.\n\nDoes anyone knows any rule in order to compose the submission file in perfect order for this case?",
    "1809100": "Nice spot! If the content of the cells is identical, then I guess we're stuck, and the relative order of these particular subsets of cells will just have to be written off as noise.",
    "1813752": "In real life, this can happen. I mean, I can put headlines like \"EDA\" as markdowns in different parts of the notebook. Whether we could mark them as the noise completely depends on the frequency of the cells in the dataset."
  },
  "source": "meta"
}