Background

Currently, when Hive schema changes the table is dropped and re-created with new schema.
But this requires re-scanning of all partitions which is slow, especially on S3.

A better solution could be based on:

ALTER TABLE my_ext_table REPLACE COLUMNS (
  col1 INT,
  col2 STRING
);

Feature

When replacing Hive schema use 'ALTER TABLE' to avoid 'MSCK REPAIR TABLE'.

Example [Optional]

--

Proposed Solution [Optional]

--