{
  "id": 23279,
  "title": "part type",
  "url": "/competitions/bosch-production-line-performance/discussion/23279",
  "author_name": "",
  "post_date": "2016-09-02T07:37:44.827Z",
  "votes": null,
  "comment_count": 3,
  "views": 515,
  "content": "<p>I found that all parts (1183748) may distribute into 17144 type  </p>",
  "messages": [
    {
      "id": "133439",
      "postDate": "09/02/2016 07:37:44",
      "content": "<p>I found that all parts (1183748) may distribute into 17144 type  </p>",
      "rawMarkdown": "I found that all parts (1183748) may distribute into 17144 type",
      "votes": null
    },
    {
      "id": "133445",
      "postDate": "09/02/2016 08:35:18",
      "content": "<p>I guess that you talk about 17144 different &quot;paths&quot; among stations. That is what I found also but I am not sure if some stations do the same job, so both &quot;paths&quot; would mean same part type. And I found that many of the types have only 1 part assigned, did you find the same? </p>",
      "rawMarkdown": "I guess that you talk about 17144 different \"paths\" among stations. That is what I found also but I am not sure if some stations do the same job, so both \"paths\" would mean same part type. And I found that many of the types have only 1 part assigned, did you find the same?",
      "votes": null
    },
    {
      "id": "133501",
      "postDate": "09/02/2016 15:37:50",
      "content": "<p>Edit: \nGiantQ and Rprades - could you say more about the 17144 paths you found? I have 7911 unique instances after a quick run through, which is similar to what Alexia counted with her java-spark setup.</p>\n\n<p><a href=\"https://www.kaggle.com/c/bosch-production-line-performance/forums/t/23301/discover-data-with-java-spark\">https://www.kaggle.com/c/bosch-production-line-performance/forums/t/23301/discover-data-with-java-spark</a></p>",
      "rawMarkdown": "Edit: \r\nGiantQ and Rprades - could you say more about the 17144 paths you found? I have 7911 unique instances after a quick run through, which is similar to what Alexia counted with her java-spark setup.\r\n\r\nhttps://www.kaggle.com/c/bosch-production-line-performance/forums/t/23301/discover-data-with-java-spark",
      "votes": null
    },
    {
      "id": "134082",
      "postDate": "09/05/2016 01:18:06",
      "content": "<p>I just assumed that the parts belonged to same type  have same sequence simply.  But, as Rprades said, I am not sure some stations do the same job .And you could use spark-shell. It makes you explore data interactively in scala. And you could load data to hive, too.</p>",
      "rawMarkdown": "I just assumed that the parts belonged to same type  have same sequence simply.  But, as Rprades said, I am not sure some stations do the same job .And you could use spark-shell. It makes you explore data interactively in scala. And you could load data to hive, too.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 133445,
      "author_name": "rprades",
      "author_url": "",
      "post_date": "09/02/2016 08:35:18",
      "content": "<p>I guess that you talk about 17144 different &quot;paths&quot; among stations. That is what I found also but I am not sure if some stations do the same job, so both &quot;paths&quot; would mean same part type. And I found that many of the types have only 1 part assigned, did you find the same? </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 133501,
      "author_name": "jpmiller",
      "author_url": "",
      "post_date": "09/02/2016 15:37:50",
      "content": "<p>Edit: \nGiantQ and Rprades - could you say more about the 17144 paths you found? I have 7911 unique instances after a quick run through, which is similar to what Alexia counted with her java-spark setup.</p>\n\n<p><a href=\"https://www.kaggle.com/c/bosch-production-line-performance/forums/t/23301/discover-data-with-java-spark\">https://www.kaggle.com/c/bosch-production-line-performance/forums/t/23301/discover-data-with-java-spark</a></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 134082,
      "author_name": "giantq",
      "author_url": "",
      "post_date": "09/05/2016 01:18:06",
      "content": "<p>I just assumed that the parts belonged to same type  have same sequence simply.  But, as Rprades said, I am not sure some stations do the same job .And you could use spark-shell. It makes you explore data interactively in scala. And you could load data to hive, too.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "133439": "I found that all parts (1183748) may distribute into 17144 type",
    "133445": "I guess that you talk about 17144 different \"paths\" among stations. That is what I found also but I am not sure if some stations do the same job, so both \"paths\" would mean same part type. And I found that many of the types have only 1 part assigned, did you find the same?",
    "133501": "Edit: \r\nGiantQ and Rprades - could you say more about the 17144 paths you found? I have 7911 unique instances after a quick run through, which is similar to what Alexia counted with her java-spark setup.\r\n\r\nhttps://www.kaggle.com/c/bosch-production-line-performance/forums/t/23301/discover-data-with-java-spark",
    "134082": "I just assumed that the parts belonged to same type  have same sequence simply.  But, as Rprades said, I am not sure some stations do the same job .And you could use spark-shell. It makes you explore data interactively in scala. And you could load data to hive, too."
  },
  "source": "meta"
}