[ https://issues.apache.org/jira/browse/SPARK-27296?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16804176#comment-16804176 ]
Erik Erlandson commented on SPARK-27296: ---------------------------------------- My initial proposal would be to alter the logic underneath {code:java} register(name: String, udaf: UserDefinedAggregateFunction){code} so that the UDAF gets hooked to a TypedImperativeAggregate, and registered in the same way that objects like CountMinSketchAgg are. > User Defined Aggregating Functions (UDAFs) have a major efficiency problem > -------------------------------------------------------------------------- > > Key: SPARK-27296 > URL: https://issues.apache.org/jira/browse/SPARK-27296 > Project: Spark > Issue Type: Bug > Components: Spark Core, SQL, Structured Streaming > Affects Versions: 2.3.3, 2.4.0, 3.0.0 > Reporter: Erik Erlandson > Priority: Major > Labels: performance, usability > > Spark's UDAFs appear to be serializing and de-serializing to/from the > MutableAggregationBuffer for each row. This gist shows a small reproducing > UDAF and a spark shell session: > [https://gist.github.com/erikerlandson/3c4d8c6345d1521d89e0d894a423046f] > The UDAF and its compantion UDT are designed to count the number of times > that ser/de is invoked for the aggregator. The spark shell session > demonstrates that it is executing ser/de on every row of the data frame. > Note, Spark's pre-defined aggregators do not have this problem, as they are > based on an internal aggregating trait that does the correct thing and only > calls ser/de at points such as partition boundaries, presenting final > results, etc. > This is a major problem for UDAFs, as it means that every UDAF is doing a > massive amount of unnecessary work per row, including but not limited to Row > object allocations. For a more realistic UDAF having its own non trivial > internal structure it is obviously that much worse. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org