Alexander Faxå created BEAM-10314:
-------------------------------------
Summary: Writing containers containing nested Decimal to BigQuery
fails
Key: BEAM-10314
URL: https://issues.apache.org/jira/browse/BEAM-10314
Project: Beam
Issue Type: Bug
Components: io-py-gcp
Reporter: Alexander Faxå
This piece of code:
[https://github.com/apache/beam/blob/master/sdks/python/apache_beam/io/gcp/bigquery_tools.py#L946]
is attempting to handle the case when inserting objects of type
decimal.Decimal into BigQuery by converting them to string.
This works but doesn't handle the case when the `v` is really a container of
nested objects (e.g. a dict), and one of them is a decimal.Decimal. This can
easily happen when using ReadFromBigQuery on a table with nested schema.
Note how `to_json_value`, which is being called here, is recursive for this
reason. I verified that a fix for my crash is to insert
```
if isinstance(obj, decimal.Decimal):
obj = str(obj)
```
in the beginning of to_json_value.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)