[BUG] SQL/Task Aggregated Metrics contain duplicated executor CPU time results

[cindyyuanjiang]

In Profiler tool, when aggregating SQL/Task metrics results in function aggregateSparkMetricsBySql, execCpuTime is defined and used twice in SQLTaskAggMetricsProfileResult:

• spark-rapids-tools/core/src/main/scala/com/nvidia/spark/rapids/tool/analysis/AppSparkMetricsAnalyzer.scala

  Line 212 in 9564d0b

  execCpuTime,

• spark-rapids-tools/core/src/main/scala/com/nvidia/spark/rapids/tool/analysis/AppSparkMetricsAnalyzer.scala

  Line 220 in 9564d0b

  execCpuTime,

We should remove this duplicated metric result.

[amahussein]

Yes, I noticed that too when I was doing the optimizations in #1461
There should be only execCpuTimeSum because SQL does not have execCpuTime. It has to be aggregated on the tasks.