Skip to content

Fix list feature format for BigQuery offline datasources. #1889

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
feast-ci-bot merged 7 commits into from
Sep 21, 2021
Merged

Fix list feature format for BigQuery offline datasources. #1889

Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
fdf5722
Enable test for list features in BQ
judahrand Sep 20, 2021
268075e
Remove specifc handling of BQ list types
judahrand Sep 20, 2021
6279b9d
Enable Parquet list inference for BigQuery
judahrand Sep 20, 2021
5d35ec5
Enable `use_compliant_nested_type`
judahrand Sep 20, 2021
81c8abc
Add potentially missing import
judahrand Sep 20, 2021
db6f607
Upload all data to BQ in ARRAY safe manner
judahrand Sep 20, 2021
f90aa17
Handle empty list in `python_value_to_proto_value`
judahrand Sep 20, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Enable use_compliant_nested_type 
Signed-off-by: Judah Rand <17158624+judahrand@users.noreply.github.com>
  • Loading branch information
@judahrand
judahrand committed Sep 20, 2021
commit 5d35ec5942da4f9d469faf09434b3dbe328d601b
26 changes: 20 additions & 6 deletions sdk/python/feast/infra/offline_stores/bigquery.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -222,15 +222,23 @@ def to_bigquery(
job_config = bigquery.QueryJobConfig(destination=path)

if not job_config.dry_run and self.on_demand_feature_views is not None:
transformed_df = self.to_df()
# It is complicated to get BQ to understand that we want an ARRAY<value_type>
# https://cloud.google.com/bigquery/docs/reference/rest/v2/tables#parquetoptions
# https://github.com/googleapis/python-bigquery/issues/19
writer = pyarrow.BufferOutputStream()
pyarrow.parquet.write_table(
self.to_arrow(), writer, use_compliant_nested_type=True
)
reader = pyarrow.BufferReader(writer.getvalue())

parquet_options = bigquery.format_options.ParquetOptions()
parquet_options.enable_list_inference = True
job_config = bigquery.LoadJobConfig()
job_config.source_format = bigquery.SourceFormat.PARQUET
job_config.parquet_options = parquet_options
job = self.client.load_table_from_dataframe(
transformed_df, job_config.destination, job_config=job_config,

job = self.client.load_table_from_file(
reader, job_config.destination, job_config=job_config,
)
job.result()
print(f"Done writing to '{job_config.destination}'.")
Expand Down Expand Up @@ -337,15 +345,21 @@ def _upload_entity_df_and_get_entity_schema(
entity_df.reset_index(drop=True, inplace=True)

# Upload the dataframe into BigQuery, creating a temporary table
# It is complicated to get BQ to understand that we want an ARRAY<value_type>
# https://cloud.google.com/bigquery/docs/reference/rest/v2/tables#parquetoptions
# https://github.com/googleapis/python-bigquery/issues/19
writer = pyarrow.BufferOutputStream()
pyarrow.parquet.write_table(
pyarrow.Table.from_pandas(entity_df), writer, use_compliant_nested_type=True
)
reader = pyarrow.BufferReader(writer.getvalue())

parquet_options = bigquery.format_options.ParquetOptions()
parquet_options.enable_list_inference = True
job_config = bigquery.LoadJobConfig()
job_config.source_format = bigquery.SourceFormat.PARQUET
job_config.parquet_options = parquet_options
job = client.load_table_from_dataframe(
entity_df, table_name, job_config=job_config
)
job = client.load_table_from_file(reader, table_name, job_config=job_config)
block_until_done(client, job)

entity_schema = dict(zip(entity_df.columns, entity_df.dtypes))
Expand Down