CODEWITHSUNDEEP
google-bigquery
Gil Jaysmith
Gil Jaysmith

Reputation: 23

bq mk with --time_partitioning_field tells me "Field based partitioning support is not yet available for this project"

I saw in the documentation for partitioning that you can partition a table based on a timestamp field in the schema, rather than on the data's insertion time. I was hoping to experiment with this by taking one of our existing tables, exporting its data, and then creating a new table with the same schema and with partitioning on the timestamp field, but when I try it I get:

"Field based partitioning support is not yet available for this project"

Is this something I have to ask to be set up for my project, or is it experimental? If the latter, is there an ETA for it being rolled?

The situation is that I have terabytes of data stored in nonpartitioned tables, and it seems like not only will the conversion process will be painful (I've read Migrating from non-partitioned to Partitioned tables), but my Dataflow pipeline going forward will have to do ugly things to write new data into the correct partition, because 'time of insertion' won't be accurate compared to the timestamps in the actual data.

Upvotes: 0

Views: 793

Answers (1)

Hua Zhang
Hua Zhang

Reputation: 1551

I guess you read about the new feature in the API references. We're preparing to alpha this feature so have enabled it in the API and client. You can track the feature progress at https://issuetracker.google.com/issues/65440943. Thanks!

Upvotes: 3

Related Questions