{
  "id": 39936,
  "title": "level3 category names are not unique",
  "url": "/competitions/cdiscount-image-classification-challenge/discussion/39936",
  "author_name": "",
  "post_date": "2017-09-24T11:17:16.093119800Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>I have a question regarding the mapping of category_id to name.\nWhile most of the names have a unique relation to the category_id, the following category names are not uniqe:\nCONFORT URINAIRE\nFONTAINE A EAU\nFUSIBLE\nGUIDON\nPELUCHE\nPROTEGE ECRAN - FILM DE PROTECTION\nVOITURE</p>\n\n<p>They have two category_id's. Is this an error in the data? Does this mean, the category_id is level 4 of the categories?</p>",
  "messages": [
    {
      "id": "223940",
      "postDate": "09/24/2017 11:17:16",
      "content": "<p>I have a question regarding the mapping of category_id to name.\nWhile most of the names have a unique relation to the category_id, the following category names are not uniqe:\nCONFORT URINAIRE\nFONTAINE A EAU\nFUSIBLE\nGUIDON\nPELUCHE\nPROTEGE ECRAN - FILM DE PROTECTION\nVOITURE</p>\n\n<p>They have two category_id's. Is this an error in the data? Does this mean, the category_id is level 4 of the categories?</p>",
      "rawMarkdown": "I have a question regarding the mapping of category_id to name.\nWhile most of the names have a unique relation to the category_id, the following category names are not uniqe:\nCONFORT URINAIRE\nFONTAINE A EAU\nFUSIBLE\nGUIDON\nPELUCHE\nPROTEGE ECRAN - FILM DE PROTECTION\nVOITURE\n\nThey have two category_id's. Is this an error in the data? Does this mean, the category_id is level 4 of the categories?",
      "votes": null
    },
    {
      "id": "226508",
      "postDate": "10/02/2017 14:39:59",
      "content": "<p>The dataset is ok you, are just narrowing the context too much for your count - if you look at level3 categories only, some have 2 rows with 2 unique category_id each. If on the other hand you aggregate on level2 &amp; level3 you will find that the combination of this 2 dimensions has always only one category_id ;)  </p>",
      "rawMarkdown": "The dataset is ok you, are just narrowing the context too much for your count - if you look at level3 categories only, some have 2 rows with 2 unique category_id each. If on the other hand you aggregate on level2 &amp; level3 you will find that the combination of this 2 dimensions has always only one category_id ;)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 226508,
      "author_name": "boryssobiegraj",
      "author_url": "",
      "post_date": "10/02/2017 14:39:59",
      "content": "<p>The dataset is ok you, are just narrowing the context too much for your count - if you look at level3 categories only, some have 2 rows with 2 unique category_id each. If on the other hand you aggregate on level2 &amp; level3 you will find that the combination of this 2 dimensions has always only one category_id ;)  </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "223940": "I have a question regarding the mapping of category_id to name.\nWhile most of the names have a unique relation to the category_id, the following category names are not uniqe:\nCONFORT URINAIRE\nFONTAINE A EAU\nFUSIBLE\nGUIDON\nPELUCHE\nPROTEGE ECRAN - FILM DE PROTECTION\nVOITURE\n\nThey have two category_id's. Is this an error in the data? Does this mean, the category_id is level 4 of the categories?",
    "226508": "The dataset is ok you, are just narrowing the context too much for your count - if you look at level3 categories only, some have 2 rows with 2 unique category_id each. If on the other hand you aggregate on level2 &amp; level3 you will find that the combination of this 2 dimensions has always only one category_id ;)"
  },
  "source": "meta"
}