Hi Bart,

I didn't try this before, because what I'm interested in is seeing the 
intermediate steps' output.

However, with the Beam-Direct runner, the pipeline just hangs:

2022/07/19 14:13:34 - Hop - Pipeline opened.
2022/07/19 14:13:34 - Hop - Launching pipeline [sites]...
2022/07/19 14:13:34 - Hop - Started the pipeline execution.
2022/07/19 14:13:41 - General - Created Apache Beam pipeline with name 'sites'
2022/07/19 14:13:41 - General - Handled generic transform (TRANSFORM) : Get 
file names, gets data from 0 previous transform(s), targets=0, infos=0
2022/07/19 14:13:41 - General - Handled generic transform (TRANSFORM) : Last 
modified, gets data from 1 previous transform(s), targets=0, infos=0
2022/07/19 14:13:41 - General - Handled generic transform (TRANSFORM) : Avro 
File Input, gets data from 1 previous transform(s), targets=0, infos=0
2022/07/19 14:13:41 - General - Handled generic transform (TRANSFORM) : Avro to 
site, gets data from 1 previous transform(s), targets=0, infos=0
2022/07/19 14:13:41 - General - Handled generic transform (TRANSFORM) : Filter 
rows, gets data from 1 previous transform(s), targets=1, infos=0
2022/07/19 14:13:41 - General - Transform Select values reading from previous 
transform targeting this one using : Filter rows - TARGET - Select values
2022/07/19 14:13:41 - General - Handled generic transform (TRANSFORM) : Select 
values, gets data from 1 previous transform(s), targets=0, infos=0
2022/07/19 14:13:42 - General - Handled transform (BQ OUTPUT) : Beam BigQuery 
Output, gets data from Select values
2022/07/19 14:13:42 - sites - Executing this pipeline using the Beam Pipeline 
Engine with run configuration 'Beam-Direct'
… nothing more happens

If I remove the "Beam BigQuery Output" transform:

2022/07/19 14:14:32 - Hop - Pipeline opened.
2022/07/19 14:14:32 - Hop - Launching pipeline [sites]...
2022/07/19 14:14:32 - Hop - Started the pipeline execution.
2022/07/19 14:14:40 - General - Created Apache Beam pipeline with name 'sites'
2022/07/19 14:14:40 - General - Handled generic transform (TRANSFORM) : Get 
file names, gets data from 0 previous transform(s), targets=0, infos=0
2022/07/19 14:14:40 - General - Handled generic transform (TRANSFORM) : Last 
modified, gets data from 1 previous transform(s), targets=0, infos=0
2022/07/19 14:14:40 - General - Handled generic transform (TRANSFORM) : Avro 
File Input, gets data from 1 previous transform(s), targets=0, infos=0
2022/07/19 14:14:40 - General - Handled generic transform (TRANSFORM) : Avro to 
site, gets data from 1 previous transform(s), targets=0, infos=0
2022/07/19 14:14:40 - General - Handled generic transform (TRANSFORM) : Filter 
rows, gets data from 1 previous transform(s), targets=1, infos=0
2022/07/19 14:14:40 - General - Transform Select values reading from previous 
transform targeting this one using : Filter rows - TARGET - Select values
2022/07/19 14:14:40 - General - Handled generic transform (TRANSFORM) : Select 
values, gets data from 1 previous transform(s), targets=0, infos=0
2022/07/19 14:14:40 - sites - Executing this pipeline using the Beam Pipeline 
Engine with run configuration 'Beam-Direct'
2022/07/19 14:14:44 - sites - Beam pipeline execution has finished.

The pipeline ran successfully via the DataFlow runner.

Fabian


> Am 19.07.2022 um 13:11 schrieb Bart Maertens <[email protected]>:
> 
> Hi Fabian,
> 
> Do you have this issue with the BeaM Direct run configuration as well? 
> The Beam Bigquery Output transform is Beam only, so this won't work with the 
> native (local or remote) run configuration. 
> 
> If the issue exists with the direct runner, can you share any errors you get? 
> 
> Regards,
> Bart 
> 
> On Tue, Jul 19, 2022 at 12:12 PM Fabian Peters <[email protected] 
> <mailto:[email protected]>> wrote:
> Hi all!
> 
> I'm developing a number of pipelines that write data to BigQuery. This works 
> fine, alas, during development I find I have to entirely remove the "Beam 
> BigQuery Output" transform to be able to run the pipeline locally, disabling 
> or deleting the hop to it did not help. Is there a way to keep the transform 
> around while debugging the pipeline?
> 
> cheers
> 
> Fabian

Reply via email to