ESQL: 'ValuesSources are mismatched' stats group by field that is missing in some indices

[craigtaverner]

Similar to the bug reported in #100186, it seems that stats over multiple indices where some might be missing the column generate errors:

FROM logs-* | STATS count=count(*) BY agent.version

Results in:

ValuesSources are mismatched

This happens in benchmarking tests where very little data is generated. The full benchmarks have all data generated, and the error does not occur, so we presume the issue might be either:

• A mismatch between documents created and index mappings
• Or missing data in indexes that do have this field mapped

The error trace is:

java.lang.IllegalStateException: ValuesSources are mismatched
	at org.elasticsearch.compute.operator.OrdinalsGroupingOperator.<init>(OrdinalsGroupingOperator.java:102)
	at org.elasticsearch.compute.operator.OrdinalsGroupingOperator.get(OrdinalsGroupingOperator.java:64)
	at org.elasticsearch.xpack.esql.planner.LocalExecutionPlanner.lambda-zsh(LocalExecutionPlanner.java:646)
	at java.base/java.util.stream.ReferencePipeline.accept(ReferencePipeline.java:197)
	at java.base/java.util.ArrayList.forEachRemaining(ArrayList.java:1625)
	at java.base/java.util.stream.AbstractPipeline.copyInto(AbstractPipeline.java:509)
	at java.base/java.util.stream.AbstractPipeline.wrapAndCopyInto(AbstractPipeline.java:499)
	at java.base/java.util.stream.ForEachOps.evaluateSequential(ForEachOps.java:150)
	at java.base/java.util.stream.ForEachOps.evaluateSequential(ForEachOps.java:173)
	at java.base/java.util.stream.AbstractPipeline.evaluate(AbstractPipeline.java:234)
	at java.base/java.util.stream.ReferencePipeline.forEach(ReferencePipeline.java:596)
	at org.elasticsearch.xpack.esql.planner.LocalExecutionPlanner.operators(LocalExecutionPlanner.java:646)
	at org.elasticsearch.xpack.esql.planner.LocalExecutionPlanner.apply(LocalExecutionPlanner.java:730)
	at org.elasticsearch.xpack.esql.planner.LocalExecutionPlanner.createDrivers(LocalExecutionPlanner.java:773)
	at org.elasticsearch.xpack.esql.plugin.ComputeService.runCompute(ComputeService.java:265)
	at org.elasticsearch.xpack.esql.plugin.ComputeService.lambda(ComputeService.java:432)
	at [email protected]/org.elasticsearch.action.ActionListener.onResponse(ActionListener.java:177)
	at [email protected]/org.elasticsearch.action.ActionListener.completeWith(ActionListener.java:305)
	at org.elasticsearch.xpack.esql.plugin.ComputeService.lambda(ComputeService.java:301)
	at [email protected]/org.elasticsearch.index.shard.IndexShard.ensureShardSearchActive(IndexShard.java:3932)
	at org.elasticsearch.xpack.esql.plugin.ComputeService.acquireSearchContexts(ComputeService.java:299)
	at org.elasticsearch.xpack.esql.plugin.ComputeService.messageReceived(ComputeService.java:430)
	at org.elasticsearch.xpack.esql.plugin.ComputeService.messageReceived(ComputeService.java:422)
	at [email protected]/org.elasticsearch.transport.RequestHandlerRegistry.processMessageReceived(RequestHandlerRegistry.java:75)
	at [email protected]/org.elasticsearch.transport.TransportService.doRun(TransportService.java:1020)
	at [email protected]/org.elasticsearch.common.util.concurrent.ThreadContext.doRun(ThreadContext.java:983)
	at [email protected]/org.elasticsearch.common.util.concurrent.AbstractRunnable.run(AbstractRunnable.java:26)
	at [email protected]/org.elasticsearch.common.util.concurrent.TimedRunnable.doRun(TimedRunnable.java:33)
	at [email protected]/org.elasticsearch.common.util.concurrent.ThreadContext.doRun(ThreadContext.java:983)
	at [email protected]/org.elasticsearch.common.util.concurrent.AbstractRunnable.run(AbstractRunnable.java:26)
	at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1144)
	at java.base/java.util.concurrent.ThreadPoolExecutor.run(ThreadPoolExecutor.java:642)
	at java.base/java.lang.Thread.run(Thread.java:1623)

[elasticsearchmachine]

Pinging @elastic/es-ql (Team:QL)

[elasticsearchmachine]

Pinging @elastic/elasticsearch-esql (:Query Languages/ES|QL)

[craigtaverner]

We needed to tweak the benchmark tests to include more data to avoid this bug. We increased the dataset time range form 2s to 1min, a 30x increase. This is a 30x increase, although only increased the total test time from about 45min to about 54min, but we've been asked to decrease this again once the bug is fixed.

[dnhatn]

I've opened #100566.