[GitHub] [druid] clintropolis commented on pull request #10921: fix 'is not null' is invalid for array-type in druid-sql

[GitBox <gi...@apache.org>]

clintropolis commented on pull request #10921:
URL: https://github.com/apache/druid/pull/10921#issuecomment-788694325


   >Hey, i had try to add a test in https://github.com/apache/druid/blob/master/sql/src/test/java/org/apache/druid/sql/calcite/CalciteQueryTest.java.
   However, I can not simulate a filter of string_array column,because we treat string_array column as string column in the org.apache.druid.sql.calcite.schema.DruidSchema.analysisToRowSignature(). For example, dim3 column in table 'foo' is a multivalue column, and in the schema ,it convert to a string column simply.
   
   Hmm, yeah, true array types aren't quite fully supported throughout the query engine yet. We've got some functions to interact with multi-value string dimensions as if they were arrays, as well as array constants, but the result row signatures are still for the most part non-array types because it's not quite there yet (no grouping on the rows as complete arrays for example). You might be able to use an [array constant on a scan query to get the right row signature](https://github.com/apache/druid/blob/master/sql/src/test/java/org/apache/druid/sql/calcite/CalciteQueryTest.java#L15055), but I'm not certain about that because in the native druid query, array types will still currently end up as STRING virtual columns I believe.


----------------------------------------------------------------
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:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org