{
  "id": 22621,
  "title": "Row means in H2o in R",
  "url": "/competitions/grupo-bimbo-inventory-demand/discussion/22621",
  "author_name": "",
  "post_date": "2016-08-02T08:25:09.203Z",
  "votes": null,
  "comment_count": 1,
  "views": 414,
  "content": "<p>Since the dataset is very large and I only own a 2gb ram laptop, I manipulated the dataset in H2o environment. I reshaped the dataset so that there are demand columns for week3, week4 .. up to week9. I wanted to get the mean demand per row but I am having a hard time finding the solution in H2o environment. Can anyone suggest a possible solution or alternative?</p>",
  "messages": [
    {
      "id": "129774",
      "postDate": "08/02/2016 08:25:09",
      "content": "<p>Since the dataset is very large and I only own a 2gb ram laptop, I manipulated the dataset in H2o environment. I reshaped the dataset so that there are demand columns for week3, week4 .. up to week9. I wanted to get the mean demand per row but I am having a hard time finding the solution in H2o environment. Can anyone suggest a possible solution or alternative?</p>",
      "rawMarkdown": "Since the dataset is very large and I only own a 2gb ram laptop, I manipulated the dataset in H2o environment. I reshaped the dataset so that there are demand columns for week3, week4 .. up to week9. I wanted to get the mean demand per row but I am having a hard time finding the solution in H2o environment. Can anyone suggest a possible solution or alternative?",
      "votes": null
    },
    {
      "id": "130012",
      "postDate": "08/03/2016 10:49:14",
      "content": "<p>I don't think you can in H2o (you can't manipulate much in H2o anyway, especially if you intend to engineer features this is not possible).</p>\n\n<p>Are you able to load the data in SQL and edit in SQL? This should be feasible in a 2GB RAM laptop. Then, you just need to pick back the database as CSV and import it back into H2o.</p>\n\n<p>P.S: I recommend you to try on a very small sample size to check and test (if it is what you expected to see), and not try with the whole dataset as a first try (because it might take a while).</p>",
      "rawMarkdown": "I don't think you can in H2o (you can't manipulate much in H2o anyway, especially if you intend to engineer features this is not possible).\r\n\r\nAre you able to load the data in SQL and edit in SQL? This should be feasible in a 2GB RAM laptop. Then, you just need to pick back the database as CSV and import it back into H2o.\r\n\r\nP.S: I recommend you to try on a very small sample size to check and test (if it is what you expected to see), and not try with the whole dataset as a first try (because it might take a while).",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 130012,
      "author_name": "laurae2",
      "author_url": "",
      "post_date": "08/03/2016 10:49:14",
      "content": "<p>I don't think you can in H2o (you can't manipulate much in H2o anyway, especially if you intend to engineer features this is not possible).</p>\n\n<p>Are you able to load the data in SQL and edit in SQL? This should be feasible in a 2GB RAM laptop. Then, you just need to pick back the database as CSV and import it back into H2o.</p>\n\n<p>P.S: I recommend you to try on a very small sample size to check and test (if it is what you expected to see), and not try with the whole dataset as a first try (because it might take a while).</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "129774": "Since the dataset is very large and I only own a 2gb ram laptop, I manipulated the dataset in H2o environment. I reshaped the dataset so that there are demand columns for week3, week4 .. up to week9. I wanted to get the mean demand per row but I am having a hard time finding the solution in H2o environment. Can anyone suggest a possible solution or alternative?",
    "130012": "I don't think you can in H2o (you can't manipulate much in H2o anyway, especially if you intend to engineer features this is not possible).\r\n\r\nAre you able to load the data in SQL and edit in SQL? This should be feasible in a 2GB RAM laptop. Then, you just need to pick back the database as CSV and import it back into H2o.\r\n\r\nP.S: I recommend you to try on a very small sample size to check and test (if it is what you expected to see), and not try with the whole dataset as a first try (because it might take a while)."
  },
  "source": "meta"
}