[Bug] Row field descriptions are ignored when creating tables with Flink SQL

### Search before asking

- [X] I searched in the [issues](https://github.com/apache/paimon/issues) and found nothing similar.

### Paimon version

a389af62b7a254ffabb91a9e43dfe8a90695e899

### Compute Engine

Flink

### Minimal reproduce step

Use `sql-client.sh` and create this table in a Paimon catalog:
```
CREATE TABLE my_table (my_row ROW);

SHOW CREATE TABLE my_table;
CREATE TABLE `my_catalog`.`my_database`.`my_table` (
`my_row` ROW<`my_field` INT>
) WITH (
'path' = 'file:/paimon_data/my_database.db/my_table'
)
```

### What doesn't meet your expectations?

Descriptions of row fields are not persisted when using Flink SQL. This is a problem when manually creating tables for the [KafkaSyncTableAction](https://paimon.apache.org/docs/master/api/java/org/apache/paimon/flink/action/cdc/kafka/KafkaSyncTableAction) which uses descriptions on row fields, causing schema mismatches.

### Anything else?

_No response_

### Are you willing to submit a PR?

- [ ] I'm willing to submit a PR!

Reproduce the issue with Flink's sql-client.sh using the CREATE TABLE statement in the report, then inspect the Flink SQL schema conversion and the SHOW CREATE TABLE path. Compare the row-field description before and after table creation, and confirm the fix preserves it so KafkaSyncTableAction no longer encounters schema mismatches.

Written by the indexing model from the issue text.