iemejia commented on code in PR #56479:
URL: https://github.com/apache/spark/pull/56479#discussion_r3442157299


##########
.github/workflows/benchmark.yml:
##########
@@ -73,7 +78,11 @@ jobs:
   # Any TPC-DS related updates on this job need to be applied to tpcds-1g job 
of build_and_test.yml as well
   tpcds-1g-gen:
     name: "Generate an TPC-DS dataset with SF=1"
-    if: contains(inputs.class, 'TPCDSQueryBenchmark') || 
contains(inputs.class, 'LZ4TPCDSDataBenchmark') || contains(inputs.class, 
'ZStandardTPCDSDataBenchmark') || contains(inputs.class, '*')
+    # Only generate TPC-DS data when running TPC-DS benchmarks or all 
benchmarks (class == '*').
+    # Use contains(inputs.class, 'TPCDS') to catch both exact class names and 
globs like '*TPCDS*'.
+    # Use exact equality for '*' to avoid matching wildcard patterns like
+    # '*VectorizedDeltaReaderBenchmark' that don't need TPC-DS data.
+    if: contains(inputs.class, 'TPCDS') || inputs.class == '*'

Review Comment:
   You're right that pattern-matching heuristics will always have edge cases -- 
`org.apache.spark.sql.execution.benchmark.*` is one, and future class names 
could introduce others.
   
   How about replacing the auto-detection with an explicit boolean input 
instead?
   
   ```yaml
   generate-tpcds:
     type: boolean
     description: 'Generate TPC-DS data (required for TPC-DS benchmarks and 
all-benchmark runs)'
     required: true
     default: true
   ```
   
   The `if:` condition simplifies to just `if: inputs.generate-tpcds`. Default 
is `true` (matching the default `class: '*'`), and users running non-TPC-DS 
benchmarks can set it to `false` to skip the ~10 min generation step. No 
heuristics, no edge cases to worry about.



-- 
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]

Reply via email to