Update document Storage API Support in `Google BigQuery I/O connector` for Python SDK #26889

RyuSA · 2023-05-25T08:39:37Z

Please add a meaningful description for your change here

This pull request contains these changes below.👏

Add usage of WriteToBigQuery.Method.STORAGE_WRITE_API to the document.
Add usage of ReadFromBigQuery.Method.DIRECT_READ to the document.

Here is an example code to check the features.
https://gist.github.com/RyuSA/84968c2322771ce411e63154593b2319

context

In the original issue(#26693), we discussed about only STORAGE_WRITE_API, but I found the BigQuery read api is also supported now. I just add the usage of the read api too. If any concerns you see please let me know.

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

Mention the appropriate issue in your description (for example: addresses #123), if applicable. This will automatically add a link to the pull request in the issue. If you would like the issue to automatically close on merging the pull request, comment fixes #<ISSUE NUMBER> instead.
Update CHANGES.md with noteworthy changes.
If this contribution is large, please file an Apache Individual Contributor License Agreement.

See the Contributor Guide for more tips on how to make review process smoother.

To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md

GitHub Actions Tests Status (on master branch)

See CI.md for more information about GitHub Actions CI.

Add usage of BigQuery StorageWriteAPI in Python.

Add usage of BigQuery StorageReadAPI in Python.

github-actions · 2023-05-25T09:05:57Z

Assigning reviewers. If you would like to opt out of this review, comment assign to next reviewer:

R: @AnandInguva for label python.

Available commands:

stop reviewer notifications - opt out of the automated review tooling
remind me after tests pass - tag the comment author after tests pass
waiting on author - shift the attention set back to the author (any comment or push by the author will return the attention set to the reviewers)

The PR bot will only process comments in the main thread (not review comments).

Abacn · 2023-05-25T14:42:09Z

R: @ahmedabu98

also, please check the "PythonFormatter" and "PythonLint" results. Could format the code by running

# Run from root beam repo dir
pip install yapf==0.29.0
git diff HEAD --name-only | grep "\.py$" | xargs yapf --in-place

github-actions · 2023-05-25T14:43:33Z

Stopping reviewer notifications for this pull request: review requested by someone other than the bot, ceding control

apply lint and format

ahmedabu98 · 2023-05-26T16:30:53Z

sdks/python/apache_beam/examples/snippets/snippets.py

+  # [START model_bigqueryio_write_with_storage_write_api]
+  quotes | beam.io.WriteToBigQuery(
+      table_spec,
+      schema=table_schema,
+      method=beam.io.WriteToBigQuery.Method.STORAGE_WRITE_API)
+  # [END model_bigqueryio_write_with_storage_write_api]


Writing with storage api currently has some limitations on supported data types. This is due to the what types are currently supported at the cross-language boundary. Important to note that the python decimal.Decimal is needed to write a NUMERIC BigQuery type. Similarly, the the Beam Timestamp type is needed to write a TIMESTAMP BigQuery type.

Also, some other types are not yet supported, like "DATETIME". Might be worth mentioning some of that, here is the full allowed list:

beam/sdks/python/apache_beam/io/gcp/bigquery_tools.py

Lines 112 to 123 in 0b43074

BIGQUERY_TYPE_TO_PYTHON_TYPE = {

"STRING": str,

"BOOL": bool,

"BOOLEAN": bool,

"BYTES": bytes,

"INT64": np.int64,

"INTEGER": np.int64,

"FLOAT64": np.float64,

"FLOAT": np.float64,

"NUMERIC": decimal.Decimal,

"TIMESTAMP": apache_beam.utils.timestamp.Timestamp,

}

Thank you for letting me know. I didn't know the limitations!
I have just added these notes.👍 0af3174

#27347
Hi, @ahmedabu98 .
I tried to write 'TIMESTAMP' type, but I got an error message

java.lang.IllegalArgumentException: Input schema is not assignable to output schema. Input schema=Fields: ....... INFO 2023-07-03T09:27:24.015855Z Field{name=ts, description=, type=LOGICAL_TYPE<beam:logical_type:micros_instant:v1>, options={{}}} ....... , Output schema=Fields: ....... INFO 2023-07-03T09:27:24.016054Z Field{name=ts, description=, type=DATETIME, options={{}}} .......

I looked into beam source codes and found that input mapping and output mapping are not matched in the schema validation code). I think output mapping should be FieldType.logicalType(SqlTypes.TIMESTAMP).
Is there a way to avoid this error, or is TIMESTAMP type not supported yet?

I looked into the code again and found this.
It seems that beam:logical_type:millis_instant:v1 is translated as FieldType.DATETIME in SchemaTranslation, which is matched with the output mapping.
To use MillisInstant, LogicalType.register_logical_type(MillisInstant) should be added to the pipeline code.

Example

LogicalType.register_logical_type(MillisInstant) quotes | beam.io.WriteToBigQuery( table_spec, schema=table_schema, write_disposition=beam.io.BigQueryDisposition.WRITE_TRUNCATE, create_disposition=beam.io.BigQueryDisposition.CREATE_IF_NEEDED, method=beam.io.WriteToBigQuery.Method.STORAGE_WRITE_API, )

Thanks for looking into this @HaeSe0ng!

@Abacn do you know if it's sufficient for us to include the line LogicalType.register_logical_type(MillisInstant) in the connector? e.g. here:

beam/sdks/python/apache_beam/io/gcp/bigquery.py

Lines 2193 to 2194 in 7eaef18

else:

# Storage Write API

Or will we have to tell users to add this line themselves?

I looked into the code again and found this. It seems that beam:logical_type:millis_instant:v1 is translated as FieldType.DATETIME in SchemaTranslation, which is matched with the output mapping. To use MillisInstant, LogicalType.register_logical_type(MillisInstant) should be added to the pipeline code.

Example

LogicalType.register_logical_type(MillisInstant) quotes | beam.io.WriteToBigQuery( table_spec, schema=table_schema, write_disposition=beam.io.BigQueryDisposition.WRITE_TRUNCATE, create_disposition=beam.io.BigQueryDisposition.CREATE_IF_NEEDED, method=beam.io.WriteToBigQuery.Method.STORAGE_WRITE_API, )

Where does LogicalType.register_logical_type(MillisInstant) reference from? I'm trying to find a workaround that's usable for now since this isn't fixed in v2.49.0. My JSON schema is correctly converted to a bigquery.TableSchema instance but when passing it to apache beam it gives the error you mentioned above.

In my case the timestamp type is incorrectly translated to datetime (LOGICAL_TYPE<beam:logical_type:micros_instant:v1> ➔ DATETIME NOT NULL).

I'd love to help solve this, please let me know of anything I can help with for testing etc.

@JoeCMoore are you getting this error even when setting LogicalType.register_logical_type(MillisInstant)?

@ahmedabu98 No, this was a workaround for the bug. If you need any more details I'd be happy to help.

ahmedabu98 · 2023-05-26T16:40:29Z

Also for the STORAGE_WRITE_API use case, we should mention that an expansion service would have to be built before running the pipeline.

If running from a Beam git clone, they can build it with ./gradlew :sdks:java:io:google-cloud-platform:expansion-service:build, then the pipeline will choose that jar by default at run-time.

If they want to specify their own expansion-service, they can build it and specify with the expansion_service parameter in WriteToBigQuery

beam/sdks/python/apache_beam/io/gcp/bigquery.py

Lines 1804 to 1806 in 0b43074

    
             num_streaming_keys=DEFAULT_SHARDS_PER_DESTINATION, 
        
             expansion_service=None): 
        
           """Initialize a WriteToBigQuery transform.

Abacn · 2023-05-26T16:46:55Z

@ahmedabu98 if run a released version of beam, expansion service can be automatically downloaded and started. The only thing needed is a Java environment (to run the expansion service)

avoiding "already exists" error

avoiding "Line too long" pylint

RyuSA · 2023-05-28T10:02:28Z

As @ahmedabu98 mentioned, some of CI failed due to the service expansion error.

RuntimeError: /home/jenkins/jenkins-slave/workspace/beam_PreCommit_Python_Coverage_Commit/src/sdks/python/test-suites/tox/py38/build/srcs/sdks/java/io/google-cloud-platform/expansion-service/build/libs/beam-sdks-java-io-google-cloud-platform-expansion-service-2.49.0-SNAPSHOT.jar not found. 
Please build the server with
cd /home/jenkins/jenkins-slave/workspace/beam_PreCommit_Python_Coverage_Commit/src/sdks/python/test-suites/tox/py38/build/srcs; ./gradlew sdks:java:io:google-cloud-platform:expansion-service:build

Can I add a gradle task sdks:java:io:google-cloud-platform:expansion-service:build to sdks:python:test-suites:tox:py38:preCommitPyCoverage and pythonPreCommit ?

ahmedabu98 · 2023-06-01T09:34:55Z

Can I add a gradle task sdks:java:io:google-cloud-platform:expansion-service:build to sdks:python:test-suites:tox:py38:preCommitPyCoverage and pythonPreCommit ?

Is there another workaround? adding sdks:java:io:google-cloud-platform:expansion-service:build as a dependency will increase the build time by quite a bit (mainly due to building all of google-cloud-platform).

ahmedabu98

left some suggestions to clarify things a little

website/www/site/content/en/documentation/io/built-in/google-bigquery.md

Python SDK has some limitations on types. Co-authored-by: Ahmed Abualsaud <[email protected]>

Co-authored-by: Ahmed Abualsaud <[email protected]>

RyuSA · 2023-06-06T15:30:10Z

@ahmedabu98

Is there another workaround?

I tried to make small changes to the CI pipeline. But this is not a good approach because I should change a lot or result longer build times.

How about creating a new snippets.py.txt and adding the STORAGE_WRITE_API example codes to the file instead of snippets.py? The root cause is that the Python runtime during CI pipeline will evalute the STORAGE_WRITE_API implementation and it tries to load the jar file . I think Python runtime does not evalute the snippets.py.txt.

Pros: small changes. few side effects. build time is file.
Cons: Python runtime will verify the examples in snippets.py. We need to verify the examples in snippets.py.txt by ourselves.

move examples that requires expansion-service from snippets.py

codecov · 2023-06-06T17:09:11Z

Codecov Report

Merging #26889 (0e02469) into master (9270ee3) will increase coverage by 0.15%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master   #26889      +/-   ##
==========================================
+ Coverage   71.97%   72.13%   +0.15%     
==========================================
  Files         747      836      +89     
  Lines      101306   102091     +785     
==========================================
+ Hits        72920    73647     +727     
- Misses      26927    26985      +58     
  Partials     1459     1459

Flag	Coverage Δ
python	`81.09% <100.00%> (+0.13%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...s/python/apache_beam/examples/snippets/snippets.py	`73.18% <100.00%> (-3.53%)`	⬇️

... and 183 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

ahmedabu98 · 2023-06-15T19:54:09Z

Hey, sorry for the delay, I was on vacation for a while.

Would it be feasible to create a new model_bigqueryio for Storage API write method? Then we can add a new test (also just for Storage API) that we can trigger with the rest of the GCP xlang tests. We can add a few decorators to the new test to make sure it runs on the GCP xlang postcommits and get skipped otherwise, e.g.:

beam/sdks/python/apache_beam/io/external/xlang_bigqueryio_it_test.py

Lines 53 to 56 in 644f539

    
           @pytest.mark.uses_gcp_java_expansion_service 
        
           @unittest.skipUnless( 
        
               os.environ.get('EXPANSION_PORT'), 
        
               "EXPANSION_PORT environment var is not provided.")

This means it won't run with the Python Coverage Commit tests, but will run on PostCommit_Python_Xlang_Gcp_Direct and PostCommit_Python_Xlang_Gcp_Dataflow

Let me know what you think of this suggestion :)

This reverts commit ff777f4.

RyuSA · 2023-06-21T11:24:20Z

@ahmedabu98 Hi, I'm also sorry for my delay. I got a sick and had to fix my P1 body issue. 😭

Let me know what you think of this suggestion :)

Yes, it makes sense. 👍
I just created a new function model_bigqueryio_xlang for cross-language BigQuery source, and a new empty test test_model_bigqueryio_xlang. Do you think I correctly understand your suggestion ?

ahmedabu98

I think this should do it

sdks/python/apache_beam/examples/snippets/snippets.py

sdks/python/apache_beam/examples/snippets/snippets_test.py

Co-authored-by: Ahmed Abualsaud <[email protected]>

ahmedabu98 · 2023-06-24T05:36:09Z

Run Python_Xlang_Gcp_Direct PostCommit

ahmedabu98 · 2023-06-24T05:36:15Z

Run Python_Xlang_Gcp_Dataflow PostCommit

ahmedabu98 · 2023-06-24T06:01:07Z

Ahh looks like we're running into a validation error:

RuntimeError: java.lang.RuntimeException: java.lang.IllegalArgumentException: Input schema is not assignable to output schema.

This check is done during pipeline construction time, it compares the input data schema with the table schema... I think we can bypass this by giving it a table that doesn't exist, it will skip the validation check.

ahmedabu98 · 2023-06-24T06:06:06Z

sdks/python/apache_beam/examples/snippets/snippets.py

+    pipeline, write_project='', write_dataset='', write_table=''):
+  """Examples for cross-language BigQuery sources and sinks."""
+
+  table_spec = 'clouddataflow-readonly:samples.weather_stations'


Suggested change

table_spec = 'clouddataflow-readonly:samples.weather_stations'

table_spec = 'clouddataflow-readonly:samples.<non-existent-table>'

so everything else can stay the same, just put a table name that doesn't exist. can even use str(uuid.uuid4()) in the name to make sure

thank you, I added your suggestion. It seems it works! 👍

to avoid CI error

RyuSA · 2023-06-25T09:13:14Z

Run Python_Xlang_Gcp_Direct PostCommit

RyuSA · 2023-06-25T09:20:33Z

Run Python_Xlang_Gcp_Dataflow PostCommit

ahmedabu98

Looks like tests are running and passing now, thank you for these changes! LGTM

@Abacn for final review and merge

Abacn · 2023-07-11T17:11:20Z

Thanks! LGTM merging for now

…` for Python SDK (apache#26889) * Update BigQuery document; Python SDK(apache#26693) Add usage of BigQuery StorageWriteAPI in Python. --------- Co-authored-by: Ahmed Abualsaud <[email protected]>

RyuSA added 2 commits May 25, 2023 08:17

Update BigQuery document; Python SDK(apache#26693)

b979eb0

Add usage of BigQuery StorageWriteAPI in Python.

Update BigQuery document; Python SDK(apache#26693)

98dfb3e

Add usage of BigQuery StorageReadAPI in Python.

github-actions bot added examples python website labels May 25, 2023

github-actions bot added the Next Action: Reviewers label May 25, 2023

Fix snippet (apache#26693)

9d7bedc

apply lint and format

ahmedabu98 reviewed May 26, 2023

View reviewed changes

RyuSA added 3 commits May 28, 2023 17:48

Add notes for WriteToBigquery (apache#26693)

0af3174

Change labels in snippet (apache#26693)

ce41d2b

avoiding "already exists" error

Shorten code_sample label (apache#26693)

ec60e32

avoiding "Line too long" pylint

RyuSA force-pushed the bq-document-update branch from 399ddee to ec60e32 Compare May 28, 2023 09:13

ahmedabu98 requested changes Jun 1, 2023

View reviewed changes

website/www/site/content/en/documentation/io/built-in/google-bigquery.md Outdated Show resolved Hide resolved

website/www/site/content/en/documentation/io/built-in/google-bigquery.md Outdated Show resolved Hide resolved

RyuSA and others added 2 commits June 6, 2023 23:54

Clarify the limitation (apache#26693)

1ee151d

Python SDK has some limitations on types. Co-authored-by: Ahmed Abualsaud <[email protected]>

Simplify how to build Python SDK (apache#26693)

ef35550

Co-authored-by: Ahmed Abualsaud <[email protected]>

Split snippets into txt and py (apache#26693)

ff777f4

move examples that requires expansion-service from snippets.py

RyuSA added 3 commits June 21, 2023 12:18

Revert "Split snippets into txt and py (apache#26693)"

a313f21

This reverts commit ff777f4.

Add new function for xlang in snippet(apache#26693)

5058da6

Fix CI errors (apache#26693)

ee20836

RyuSA force-pushed the bq-document-update branch from 4467f06 to ee20836 Compare June 21, 2023 11:21

ahmedabu98 requested changes Jun 23, 2023

View reviewed changes

sdks/python/apache_beam/examples/snippets/snippets.py Outdated Show resolved Hide resolved

sdks/python/apache_beam/examples/snippets/snippets_test.py Outdated Show resolved Hide resolved

RyuSA and others added 3 commits June 24, 2023 09:19

Update sdks/python/apache_beam/examples/snippets/snippets_test.py

a955e93

Co-authored-by: Ahmed Abualsaud <[email protected]>

Update sdks/python/apache_beam/examples/snippets/snippets.py

4e084d9

Co-authored-by: Ahmed Abualsaud <[email protected]>

Apply patch for pylint

c26822c

ahmedabu98 reviewed Jun 24, 2023

View reviewed changes

Change table_spec

0e02469

to avoid CI error

ahmedabu98 approved these changes Jun 26, 2023

View reviewed changes

Abacn merged commit 6eb1b7e into apache:master Jul 11, 2023

JoeCMoore mentioned this pull request Aug 25, 2023

[Bug]: Reading from BigQuery provides inconsistent schemas #28151

Open

15 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update document Storage API Support in `Google BigQuery I/O connector` for Python SDK #26889

Update document Storage API Support in `Google BigQuery I/O connector` for Python SDK #26889

RyuSA commented May 25, 2023 •

edited

Loading

github-actions bot commented May 25, 2023

Abacn commented May 25, 2023 •

edited

Loading

github-actions bot commented May 25, 2023

ahmedabu98 May 26, 2023 •

edited

Loading

RyuSA May 28, 2023 •

edited

Loading

HaeSe0ng Jul 3, 2023 •

edited

Loading

HaeSe0ng Jul 5, 2023

ahmedabu98 Jul 6, 2023 •

edited

Loading

JoeCMoore Aug 25, 2023

ahmedabu98 Aug 29, 2023

JoeCMoore Aug 29, 2023

ahmedabu98 commented May 26, 2023

Abacn commented May 26, 2023

RyuSA commented May 28, 2023

ahmedabu98 commented Jun 1, 2023

ahmedabu98 left a comment

RyuSA commented Jun 6, 2023

codecov bot commented Jun 6, 2023 •

edited

Loading

ahmedabu98 commented Jun 15, 2023

RyuSA commented Jun 21, 2023

ahmedabu98 left a comment

ahmedabu98 commented Jun 24, 2023

ahmedabu98 commented Jun 24, 2023

ahmedabu98 commented Jun 24, 2023

ahmedabu98 Jun 24, 2023

RyuSA Jun 25, 2023

RyuSA commented Jun 25, 2023

RyuSA commented Jun 25, 2023

ahmedabu98 left a comment •

edited

Loading

Abacn commented Jul 11, 2023

	BIGQUERY_TYPE_TO_PYTHON_TYPE = {
	"STRING": str,
	"BOOL": bool,
	"BOOLEAN": bool,
	"BYTES": bytes,
	"INT64": np.int64,
	"INTEGER": np.int64,
	"FLOAT64": np.float64,
	"FLOAT": np.float64,
	"NUMERIC": decimal.Decimal,
	"TIMESTAMP": apache_beam.utils.timestamp.Timestamp,
	}

	table_spec = 'clouddataflow-readonly:samples.weather_stations'
	table_spec = 'clouddataflow-readonly:samples.<non-existent-table>'

Update document Storage API Support in Google BigQuery I/O connector for Python SDK #26889

Update document Storage API Support in Google BigQuery I/O connector for Python SDK #26889

Conversation

RyuSA commented May 25, 2023 • edited Loading

context

GitHub Actions Tests Status (on master branch)

github-actions bot commented May 25, 2023

Abacn commented May 25, 2023 • edited Loading

github-actions bot commented May 25, 2023

ahmedabu98 May 26, 2023 • edited Loading

Choose a reason for hiding this comment

RyuSA May 28, 2023 • edited Loading

Choose a reason for hiding this comment

HaeSe0ng Jul 3, 2023 • edited Loading

Choose a reason for hiding this comment

HaeSe0ng Jul 5, 2023

Choose a reason for hiding this comment

ahmedabu98 Jul 6, 2023 • edited Loading

Choose a reason for hiding this comment

JoeCMoore Aug 25, 2023

Choose a reason for hiding this comment

ahmedabu98 Aug 29, 2023

Choose a reason for hiding this comment

JoeCMoore Aug 29, 2023

Choose a reason for hiding this comment

ahmedabu98 commented May 26, 2023

Abacn commented May 26, 2023

RyuSA commented May 28, 2023

ahmedabu98 commented Jun 1, 2023

ahmedabu98 left a comment

Choose a reason for hiding this comment

RyuSA commented Jun 6, 2023

codecov bot commented Jun 6, 2023 • edited Loading

Codecov Report

ahmedabu98 commented Jun 15, 2023

RyuSA commented Jun 21, 2023

ahmedabu98 left a comment

Choose a reason for hiding this comment

ahmedabu98 commented Jun 24, 2023

ahmedabu98 commented Jun 24, 2023

ahmedabu98 commented Jun 24, 2023

ahmedabu98 Jun 24, 2023

Choose a reason for hiding this comment

RyuSA Jun 25, 2023

Choose a reason for hiding this comment

RyuSA commented Jun 25, 2023

RyuSA commented Jun 25, 2023

ahmedabu98 left a comment • edited Loading

Choose a reason for hiding this comment

Abacn commented Jul 11, 2023

Update document Storage API Support in `Google BigQuery I/O connector` for Python SDK #26889

Update document Storage API Support in `Google BigQuery I/O connector` for Python SDK #26889

RyuSA commented May 25, 2023 •

edited

Loading

Abacn commented May 25, 2023 •

edited

Loading

ahmedabu98 May 26, 2023 •

edited

Loading

RyuSA May 28, 2023 •

edited

Loading

HaeSe0ng Jul 3, 2023 •

edited

Loading

ahmedabu98 Jul 6, 2023 •

edited

Loading

codecov bot commented Jun 6, 2023 •

edited

Loading

ahmedabu98 left a comment •

edited

Loading