Clean

How Do I Determine Whether Clean Is Being Executed?

Criteria:

Check the .hoodie directory. If a metadata file whose name ends with .clean exists in the directory, the clean operation is being performed.

Example:

• The clean operation is not performed on Hudi.

  

• The clean operation is performed on Hudi.

  

How Do I Determine Whether Clean Is Properly Set?

Criteria:

Randomly find a recently written Parquet file. The content before the first underscore (_) in the Parquet file name is the file ID. The number of Parquet files with the same file ID is the number of file versions. In general scenarios, the number of Hudi table file versions cannot exceed 5 or less.

Example:

The maximum number of file versions is 2, and the clean operation is normal.