All Products
Search
Document Center

Realtime Compute for Apache Flink:SQL modifications and compatibility

Last Updated:Jun 16, 2026

Different SQL modifications affect the compatibility between a deployment and its state data in different ways.

Important

To check the compatibility between a deployment and its state data after a modification, view the state compatibility check result in the Start Job panel on the Deployments page in the development console of Realtime Compute for Apache Flink.

Scenario

Description

References

Modify a query

Changing an alias does not affect state compatibility.

Alias

Compatibility impact of modifying a GROUP BY aggregate function.

Group Aggregate

Compatibility impact of modifying an OVER aggregate function.

Over Aggregate

Compatibility impact of modifying a GROUP BY window aggregate function.

Group Window Aggregate

Compatibility impact of modifying a window table-valued function (TVF).

Window TVF

Compatibility impact of modifying deduplication logic.

Deduplication

Compatibility impact of modifying a TopN function.

Top-N

Compatibility impact of modifying a window-based TopN function.

Window Top-N

Compatibility impact of modifying a regular join, interval join, temporal join, or lookup join.

Join

Compatibility impact of modifying pattern recognition.

Pattern Recognition

Compatibility impact of modifying the WHERE clause.

Where

Compatibility impact of modifying UNION ALL.

Union All

Compatibility impact of modifying temporal sorting.

Temporal Sort

Compatibility impact of modifying the LIMIT clause.

LIMIT

Modify the source connector

Compatibility impact of modifying the source connector.

Modify a source table

Modify a sink connector

Compatibility impact of modifying a sink connector.

Change a sink connector

Other limits

Limits that apply when you modify other parts of an SQL statement, excluding the query, source connector, and sink connector.

Other limits

The following DDL statement creates the source table used in the sample SQL statements for the preceding scenarios.

CREATE TEMPORARY TABLE MyTable (
  a int,
  b bigint,
  c varchar,
  d int,
  ts timestamp(3),
  proctime as proctime(),
  watermark for ts AS ts - interval '1' second
) WITH (
...);