Seek solutions, ideas, best can give me the reference code

CodePudding user response:

Load directly with the spark in the merge,

Spark. Read. Parquet (" parquet_file_path "). Coalesce (1). Write. Parquet (" new_path ")