[jira] [Created] (HUDI-3206) Unify Hive's MOR `InputFormat` implementations (Parquet, HFile)

["Alexey Kudinkin (Jira)" <ji...@apache.org>]

Alexey Kudinkin created HUDI-3206:
-------------------------------------

             Summary: Unify Hive's MOR `InputFormat` implementations (Parquet, HFile)
                 Key: HUDI-3206
                 URL: https://issues.apache.org/jira/browse/HUDI-3206
             Project: Apache Hudi
          Issue Type: Sub-task
            Reporter: Alexey Kudinkin


Essentially, HIve's different MOR implementations should only differ in the file-format of the actual base-files written. 

 

Today, that's not the case: currently Hive's MOR `InputFormat` implementations inherit from their respective COW file-format counterparts (

`HoodieParquetInputFormat`, `HoodieHFileInputFormat`).

 

Instead we should unify both MOR impls to have common base class, which separate file-format specific impls would extend (Parquet, HFile), only overriding the `getRecordReader` method



--
This message was sent by Atlassian Jira
(v8.20.1#820001)