You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@beam.apache.org by "Ahmet Altay (JIRA)" <ji...@apache.org> on 2018/09/11 04:01:00 UTC
[jira] [Commented] (BEAM-5352) BigQuery IO Source is not Exporting
to GCS as written in documentation
[ https://issues.apache.org/jira/browse/BEAM-5352?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16610087#comment-16610087 ]
Ahmet Altay commented on BEAM-5352:
-----------------------------------
Dataflow implementation uses native code (separate from the Beam implementation) to read from BigQuery and that implementation uses GCS export followed by a parallel read as documented.
It would be nice to move this implementation to a Beam code eventually.
> BigQuery IO Source is not Exporting to GCS as written in documentation
> ----------------------------------------------------------------------
>
> Key: BEAM-5352
> URL: https://issues.apache.org/jira/browse/BEAM-5352
> Project: Beam
> Issue Type: Bug
> Components: sdk-py-core
> Affects Versions: 2.6.0
> Reporter: Rendy Bambang Junior
> Assignee: Ahmet Altay
> Priority: Major
> Labels: bigquery, gcp, python
>
> Did some check on the beam code and find out that DataFlow is querying BigQuery and retrieve the result using pagination [1]. As per our understanding, this means no parallelism on reading BigQuery table. It is contradictory to what the documentation is telling us [2].
>
> Is this some kind of work in progress? I'm filing as a bug since documentation telling me that it is using GCS meanwhile it's using NativeSourceReader which yield data per row as iterator.
>
> [1] [https://github.com/apache/beam/blob/520b3a24e49306c30940ceab09100d775a04d28e/sdks/python/apache_beam/io/gcp/bigquery.py#L1083]
> [2] [https://github.com/apache/beam/blob/520b3a24e49306c30940ceab09100d775a04d28e/sdks/python/apache_beam/io/gcp/bigquery.py#L60]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)