[Feature] Support writing CDC data in sink table's column order for StarRock

[zhuxt2015]

Purpose of the pull request

When using CDCSOURCE to sync data from a source database (e.g. PostgreSQL) to
StarRocks, the generated Flink DDL was built from the source table's schema.
If the column order in the StarRocks sink table differs from the source, data
would be written into the wrong columns.

This PR fetches the sink table's actual column metadata at job submission time
and uses it as the basis for DDL generation and INSERT statements, ensuring
data is written in the correct column order as defined in the destination table.

Brief change log

Core Logic (CreateCDCSourceOperation)

• Added setSinkTable(): matches each source table to its corresponding sink
  table by name (supports prefix/suffix/schema-prefix rules) and attaches it
  as table.sinkTable.
• Added getSinkTables(): lists all tables from the sink schema using the
  sink driver.
• Refactored getDriver(): extracted connector-specific URL resolution
  (starrocks → jdbc-url, doris → fenodes, jdbc → url) into a
  single reusable method, removing duplication with checkAndCreateSinkSchema.

Sink DDL Generation (SQLSinkBuilder)

• addSinkInsert() now resolves the effective table object as
  table.getSinkTable() when available, falling back to the source table.
  This makes both getFlinkDDL and createInsertOperations use the sink
  table's column definitions.

Model (Table)

• Added sinkTable field to carry the resolved sink-side table metadata
  alongside the source table object.

Metadata Integration (PostgresCDCBuilder)

• Implemented parseMetaDataConfig() with a helper composeJdbcProperties()
  to correctly append JDBC connection parameters (?key=value&...) to the
  PostgreSQL metadata URL.

Verify this pull request

Test class	Coverage
StarRocksTypeConvertTest	datetime→TIMESTAMP, date→DATE mappings; regression for other types and convertToDB
PostgresCDCBuilderTest	composeJdbcProperties edge cases: null / empty / single / multiple params, trailing & guard
CreateCDCSourceOperationTest	setSinkTable via reflection: plain match, schema-prefixed name, no match, driver column loading, first-match-wins
TableTest	New sinkTable field: default null, set/get, source field isolation, reset to null

1. Set up a PostgreSQL → StarRocks CDC pipeline via CDCSOURCE.
2. Create the StarRocks target table with columns in a different order than
   the source table.
3. Submit the job and confirm that data lands in the correct columns.

This pull request is code cleanup without any test coverage.

(or)

This pull request is already covered by existing tests, such as (please describe tests).

(or)

This change added tests and can be verified as follows: