[SPARK-50292] Add MapStatus RowCount optimize skewed job

[xy2953396112]

What changes were proposed in this pull request?

Optimize skew job with rowCount Metrics

Why are the changes needed?

In the production scenario, we see small differences in size, but large differences in rowCount, which cause the skew job

Does this PR introduce any user-facing change?

No

How was this patch tested?

Unit test

Was this patch authored or co-authored using generative AI tooling?

[wangyum]

Will setting the following config handle skew cases?

--conf spark.shuffle.accurateBlockSkewedFactor=10;
--conf spark.shuffle.accurateBlockThreshold=209715200;
--conf spark.shuffle.maxAccurateSkewedBlockNumber=60