0

Databrew recipes can be written under JSON for transformations that will be used more than once for multiple datasets.

This is an example that i copied from Databrew Developer Guide to do joins between datasets:

`

{
    "Action": {
        "Operation": "JOIN",
        "Parameters": {
            "joinKeys": "[{\"key\":\"assembly_session\",\"value\":\"assembly_session\"},{\"key\":\"state_code\",\"value\":\"state_code\"}]",
            "joinType": "INNER_JOIN",
            "leftColumns": "[\"year\",\"assembly_session\",\"state_code\",\"state_name\",\"all_votes\",\"yes_votes\",\"no_votes\",\"abstain\",\"idealpoint_estimate\",\"affinityscore_usa\",\"affinityscore_russia\",\"affinityscore_china\",\"affinityscore_india\",\"affinityscore_brazil\",\"affinityscore_israel\"]",
            "rightColumns": "[\"assembly_session\",\"vote_id\",\"resolution\",\"state_code\",\"state_name\",\"member\",\"vote\"]",
            "secondInputLocation": "s3://databrew-public-datasets-us-east-1/votes.csv",
            "secondaryDatasetName": "votes"
        }
    }
}

`

It's possible to select all columns with a * within "leftColumns" or anything close to that?

I've tried to add only * but it doesn't work. I will do the same transformations in multiple tables and this functionality would work great if i could select everything on a left join, without needing to specify all the columns.

tenayta
  • 1
  • 2

0 Answers0