robertwb commented on a change in pull request #12133:
URL: https://github.com/apache/beam/pull/12133#discussion_r449266867
##########
File path: sdks/python/apache_beam/transforms/sql.py
##########
@@ -67,10 +66,9 @@ class SqlTransform(ExternalTransform):
"""
URN = 'beam:external:java:sql:v1'
- def __init__(self, query, dialect=None):
+ def __init__(self, query, dialect=None, expansion_service=None):
super(SqlTransform, self).__init__(
self.URN,
NamedTupleBasedPayloadBuilder(
SqlTransformSchema(query=query, dialect=dialect)),
- BeamJarExpansionService(
- ':sdks:java:extensions:sql:expansion-service:shadowJar'))
+ expansion_service=expansion_service)
Review comment:
Yes, the `beam_services` flag is primarily for developers/testing. The
default should just work after running "pip install apache_beam" which is what
`BeamJarExpansionService` aims to provide. Externally-started job servers
running universal expansion services at a fixed port was useful for
bootstrapping, but is a bit of a special configuration. Exactly how to best
configure this (which, in large part, boils down to being able to name
expansion (or other) services), is a good open question.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
