[SparkLoad] dpp read parquet/orc optimization

[wyb]

Enhancement

Current SparkDpp will execute spark.read().orc/parquet (single_file) and sourceData.count() for each parquet file, each file will generate a job, and the jobs are serialized. If there are many files, it will take a long time.

Suggested optimizations:
read parquet files under a file group together, and need to consider the scenario of reading partition column from the path.

[wyb]

Anyone who are interested can optimize it.

[github-actions]

We have marked this issue as stale because it has been inactive for 6 months. If this issue is still relevant, removing the stale label or adding a comment will keep it active. Otherwise, we'll close it in 10 days to keep the issue queue tidy. Thank you for your contribution to StarRocks!

[MaxWk]

assign it to me pls.