sclukas77 commented on a change in pull request #12341:
URL: https://github.com/apache/beam/pull/12341#discussion_r469973877
##########
File path:
sdks/java/extensions/sql/src/main/java/org/apache/beam/sdk/extensions/sql/meta/provider/SchemaIOTableProviderWrapper.java
##########
@@ -80,13 +80,17 @@ public BeamSqlTable buildBeamSqlTable(Table
tableDefinition) {
}
}
- private BeamTableStatistics getTableStatistics(PipelineOptions options) {
+ public BeamTableStatistics getTableStatistics(PipelineOptions options) {
if (isBounded().equals(PCollection.IsBounded.BOUNDED)) {
return BeamTableStatistics.BOUNDED_UNKNOWN;
}
return BeamTableStatistics.UNBOUNDED_UNKNOWN;
}
+ public BeamTableStatistics getTableStatistics(PipelineOptions options,
SchemaIO schemaIO) {
+ return getTableStatistics(options);
+ }
+
Review comment:
The reason why I added this additional getTableStatistics() function was
that DataStoreV1 is the first IO whose getTableStatistics() function relied on
the schemaIO data, and this SchemaIOTableWrapper#getTableStatistics function
could be overridden in DataStoreV1TableProvider. The other IOs so far have not
required the schemaIO. Do you think we should get rid of the other
getTableStatistics() function and always require a schemaIO, even when the
schemaIO isn't absolutely necessary? Or support both cases?
----------------------------------------------------------------
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]
