[FEATURE] Allow Custom PARTITION BY When Using Snowflake ID (bigint) in Initial Migration

[zhongjun96]

Use Case Description

During initial migration, we noticed that the automatically created tables use the following clause:

PARTITION BY intDiv(id, 4294967)

How is the value 4294967 determined?

The Snowflake-style scheme we use (e.g., 1849360358546407424)

when performing batch inserts, this partitioning scheme creates too many partitions triggers the max_partitions_per_insert_block limit.

Since we have hundreds of tables, manually modifying each table's PARTITION BY clause is extremely tedious.

Proposed Solution

Allow users to customize the PARTITION BY clause during automatic table creation.

If a created_at or similar datetime field exists, we would prefer to partition by toYYYYMM(created_at) instead of using a division on the ID.