[Enhancement] Support Ignoring Null Values in Partial Column Updates for Unique Key Model

[hpkaiq]

Search before asking

• I had searched in the issues and found no similar issues.

Description

When using the Unique Key model with partial column update functionality, Doris should support an option to ignore null values during updates. Specifically, if a new row is inserted where the unique key already exists in the table, and the new row contains NULL values for certain columns, Doris should retain the existing values in those columns instead of overwriting them with NULL.

Example Scenario:
Existing row in table:
id = 1, name = "Alice", age = 30

New row inserted:
id = 1, name = NULL, age = 31

Expected behavior with this feature:
id = 1, name = "Alice", age = 31

(The name column retains its original value since the new row provided NULL.)

Motivation: This feature would make partial updates more flexible and reliable, especially in cases where incoming data streams may not always provide complete information. It avoids unintentionally overwriting valid existing data with NULL values, ensuring data integrity while still allowing updates to specific fields.

Solution

Prevents data loss due to NULL overwrites.
Simplifies ETL and data ingestion workflows where partial updates are common.
Provides more intuitive and user-friendly behavior for real-world use cases.

Are you willing to submit PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct