[GitHub] [spark] maropu edited a comment on issue #20965: [SPARK-21870][SQL] Split aggregation code into small functions

maropu edited a comment on issue #20965: [SPARK-21870][SQL] Split aggregation code into small functions URL: https://github.com/apache/spark/pull/20965#issuecomment-528350052 I checked the TPCDS performance numbers (sf=5) and I couldn't find much difference with/without this pr: https://d

[GitHub] [spark] maropu edited a comment on issue #20965: [SPARK-21870][SQL] Split aggregation code into small functions

maropu edited a comment on issue #20965: [SPARK-21870][SQL] Split aggregation code into small functions URL: https://github.com/apache/spark/pull/20965#issuecomment-527814216 Thanks for the check, @rednaxelafx ! As @cloud-fan said, I think that code length/bytecode mismatches are our well-

[GitHub] [spark] maropu edited a comment on issue #20965: [SPARK-21870][SQL] Split aggregation code into small functions

maropu edited a comment on issue #20965: [SPARK-21870][SQL] Split aggregation code into small functions URL: https://github.com/apache/spark/pull/20965#issuecomment-525609519 ok, I'll brush up the code more based on the comment 1, 2, and 3. > under the "generated code in the current

[GitHub] [spark] maropu edited a comment on issue #20965: [SPARK-21870][SQL] Split aggregation code into small functions

maropu edited a comment on issue #20965: [SPARK-21870][SQL] Split aggregation code into small functions URL: https://github.com/apache/spark/pull/20965#issuecomment-519743754 @gatorsmile [CodeGenerator. getLocalInputVariableValues](https://github.com/apache/spark/pull/20965/files#diff-8bcc