Reputation: 35
I have a doubleclick csv file with 20 columns (Timestamp,AdvertiserId,ActionName,Brower,OSID ...) without any header. I would like to ingest only first 3 columns into a BQ table. Is there any way to achieve that without mapping each and every column one-by-one manually into BQ's UI (create new_table ->"Schema" section)? Fields in CSV is comma separated and newlines are defined as semi-colon';'.
Upvotes: 1
Views: 1631
Reputation: 1269
There are two possible ways to do that: BigQuery: Load from CSV, skip columns
In your case I would probably suggest the second approach. Set the ignoreUnknownValues flag and pass in a schema with just the first three columns. For example:
bq load --ignore_unknown_values dataset.new_table gs://path/to/file.csv ~/path/to/schema.json
Upvotes: 1