[
https://issues.apache.org/jira/browse/BEAM-1908?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15961985#comment-15961985
]
ASF GitHub Bot commented on BEAM-1908:
--------------------------------------
GitHub user joshfg opened a pull request:
https://github.com/apache/beam/pull/2474
[BEAM-1908] Allow setting CREATE_NEVER when using a tablespec in BigQueryIO
Currently, the BigQueryIO.Write transform doesn't allow
CreateDisposition.CREATE_NEVER to be set when using a tablespec to determine
the BigQuery table name dynamically at runtime.
This check was originally put in because it was assumed that BigQueryIO
would need to create a new table every time the tablespec returns a table name
that hasn't been seen before. Since then a new BigQuery feature was released
which enables date partitioning within a single table, and is now the preferred
way to shard data by date (instead of having a table per date). It should
therefore be possible to use tablespec to write to a specific partition e.g.
`my-project:dataset.my_table$20170407`, while setting
CreateDisposition.CREATE_NEVER since we never need to create a new table.
This PR removes the check so that CreateDisposition.CREATE_NEVER can be
used with a tablespec.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/joshfg/beam beam-1908
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/beam/pull/2474.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #2474
----
commit 41dc5921891ed256632d257e412a03357476524f
Author: Josh <[email protected]>
Date: 2017-04-07T18:24:41Z
[BEAM-1908] Allow setting CREATE_NEVER when using a tablespec in BigQueryIO
----
> Allow setting CREATE_NEVER when using a tablespec in BigQueryIO
> ---------------------------------------------------------------
>
> Key: BEAM-1908
> URL: https://issues.apache.org/jira/browse/BEAM-1908
> Project: Beam
> Issue Type: Improvement
> Components: sdk-java-extensions
> Affects Versions: 0.6.0
> Reporter: Josh Forman-Gornall
> Assignee: Davor Bonaci
> Priority: Trivial
>
> Currently, the BigQueryIO.Write transform doesn't allow
> CreateDisposition.CREATE_NEVER to be set when using a tablespec to determine
> the BigQuery table name dynamically at runtime.
> This check was originally put in because it was assumed that BigQueryIO would
> need to create a new table every time the tablespec returns a table name that
> hasn't been seen before. Since then a new BigQuery feature was released which
> enables date partitioning within a single table, and is now the preferred way
> to shard data by date (instead of having a table per date). It should
> therefore be possible to use tablespec to write to a specific partition e.g.
> `my-project:dataset.my_table$20170407`, while setting
> CreateDisposition.CREATE_NEVER since we never need to create a new table.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
