Consider removing the requirement that users need to base64-encode their BYTES before passing them to BQ IO connector.
Open
bug
gcp
io
java
P3
python
- Dominant language
- Java
- Stars
- 8.7k
- Forks
- 4.7k
- Avg merge
- 1d 20h
- Merged PRs (30d)
- 196
Description
Currently, when BigQuery IO connector reads or stores BYTES datatype in BigQuery, there is an expectation that users base64-encode the bytes before passing them to the connector, see [1,2].
This may be an extra overhead for users to do base64 encoding when interacting with Beam, that is possible to avoid. Filing this issue to reconsider this behavior.
cc: [~chamikara].
[1] https://lists.apache.org/thread.html/0c2178cf8e5d9e77c4f233f05a0b87b6011a1daa1a5ae47b41463af5@%3Cdev.beam.apache.org%3E
[2] https://issues.apache.org/jira/browse/BEAM-7326
Imported from Jira [BEAM-7344](https://issues.apache.org/jira/browse/BEAM-7344). Original Jira may contain additional context.
Reported by: tvalentyn.
Contributor guide
Assessment
This issue has not been assessed yet.