gortiz commented on code in PR #12704:
URL: https://github.com/apache/pinot/pull/12704#discussion_r1582896361
##########
pinot-query-runtime/src/main/java/org/apache/pinot/query/runtime/operator/LeafStageTransferableBlockOperator.java:
##########
@@ -135,14 +160,128 @@ protected TransferableBlock getNextBlock()
}
}
- private TransferableBlock constructMetadataBlock() {
- // All data blocks have been returned. Record the stats and return EOS.
- Map<String, String> executionStats = _executionStats;
+ private void mergeExecutionStats(@Nullable Map<String, String>
executionStats) {
if (executionStats != null) {
- OperatorStats operatorStats = _opChainStats.getOperatorStats(_context,
getOperatorId());
- operatorStats.recordExecutionStats(executionStats);
+ for (Map.Entry<String, String> entry : executionStats.entrySet()) {
+ DataTable.MetadataKey key =
DataTable.MetadataKey.getByName(entry.getKey());
+ if (key == null) {
+ LOGGER.debug("Skipping unknown execution stat: {}", entry.getKey());
+ continue;
+ }
+ switch (key) {
+ case UNKNOWN:
+ LOGGER.debug("Skipping unknown execution stat: {}",
entry.getKey());
+ break;
+ case TABLE:
Review Comment:
I think this stat is very useful to indicate the table in the json stats
returned to the client. Specifically, in the json that correspond to the leaf
stage. Right now it is very difficult to map from stages to the original SQL
and this is one of the few hints we have to understand how stages are being
generated. See for example the stat map in the description of this PR. Without
the `table` attribute it is very difficult to map from these stats to the
original SQL and therefore understand how to improve your query.
Said that, we can do that without adding _table_ as an stat. We can add add
the _table_ json field on `MultiStageBrokerRequestHandler` when the json is
being created. I've pushed a change that does that.
--
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: commits-unsubscr...@pinot.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org
For additional commands, e-mail: commits-h...@pinot.apache.org
