0

I am needing to update values in a table. The table is not partitioned. New inserts are being regularly (10s to 100s per hour) streamed to the table. This results in the constant presence of a streaming buffer for this table, yielding the following message when updates are attempted:

UPDATE or DELETE statement over table <project.dataset.table>... would affect rows in the streaming buffer, which is not supported

None of the rows I am attempting to update are new and thus do not affect those inserts in the streaming buffer. Is there any way to successfully complete this update?

xibalba1
  • 538
  • 8
  • 16
  • 1
    Please carefully pick a filter that guarantees to get only old (>30minutes) data. If you still see the error, fire a bug here: https://cloud.google.com/support/docs/issue-trackers – Yun Zhang Oct 20 '19 at 01:42
  • @YunZhang, this worked. Thank you. – xibalba1 Oct 21 '19 at 14:16

1 Answers1

0

As @yun-zhang pointed out in his comment, the solution was to carefully write the WHERE clause such that only old data (>30 minutes) would be affected by the update, which solves for the issue I was experiencing.

xibalba1
  • 538
  • 8
  • 16