In a SQL Server environment, when Spring Batch uses a sharding data source, the auto-increment table of the batch framework encounters an execution error.

[liangleiA]

The tables for Spring Batch are managed by the framework itself. In the SQL Server environment, only the business tables are sharded, while the framework's BATCH_* tables are not configured with any sharding rules. However, the sharding parser modifies the default execution statements.

The table creation statement is as follows, with an auto-incrementing primary key:

CREATE TABLE BATCH_JOB_SEQ (ID BIGINT IDENTITY);

The statement is altered by the AbstractSQLBuilder, which adds extra fields to the execution, causing the SQL Server to fail.

Original SQL (executes successfully):

INSERT INTO BATCH_JOB_SEQ DEFAULT VALUES;

Transformed SQL (execution fails):

INSERT INTO BATCH_JOB_SEQ(ID, id) DEFAULT VALUES;