andygrove commented on code in PR #3076:
URL: https://github.com/apache/datafusion-comet/pull/3076#discussion_r2696020163
##########
benchmarks/pyspark/run_benchmark.py:
##########
@@ -41,9 +41,13 @@ def run_benchmark(spark: SparkSession, data_path: str, mode:
str) -> int:
start_time = time.time()
- # Repartition by a different key to force full shuffle of all columns
- # This shuffles all 50 columns including nested structs, arrays, maps
- repartitioned = df.repartition(200, "group_key")
+ # Repartition to force full shuffle of all columns
+
+ # repartition using round-robin partitioning
+ repartitioned = df.repartition(200)
+
+ # repartition using hash partitioning
+ # repartitioned = df.repartition(200, "group_key")
Review Comment:
Yes, PR https://github.com/apache/datafusion-comet/pull/3080 refactors this
to support multiple named benchmarks
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
