shardingsphere conflict with seata

When only use shardingsphere ： sharding is OK （only choise one db and one table）

When only use shardingsphere with seata ： sharding is ERROR（Scan all db and table） like
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_0 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_1 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_2 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_3 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_4 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_5 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_6 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_7 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_8 LIMIT 1
15:16:15.370 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_9 LIMIT 1
15:16:15.371 INFO [main] ShardingSphere-SQL - Actual SQL: bookshelf_0 ::: SELECT * FROM shelf_book_key_10 LIMIT 1
......
sharding-transaction-base-seata-at It doesn't help.

org.apache.shardingsphere
sharding-jdbc-spring-boot-starter
4.1.1


org.apache.shardingsphere
sharding-jdbc-spring-namespace
4.1.1

org.apache.shardingsphere
sharding-transaction-base-seata-at
4.1.2

notice ：4.1.2 is my version be used for fix SeataATShardingTransactionManager can't read seata.conf

io.seata
seata-spring-boot-starter
1.3.0

Start by reproducing the reported behavior with the listed ShardingSphere 4.1.1, Seata 1.3.0, and sharding-transaction-base-seata-at 4.1.2 dependencies. Compare the generated Actual SQL when ShardingSphere runs alone with the SQL produced when Seata is enabled. Done means the sharded query targets only the expected database and table rather than scanning every table.

Written by the indexing model from the issue text.