[Docs][Connector-V2] Improve Doris Iceberg StarRocks Paimon and Clickhouse connector docs

[DanielCarter-stack]

This PR improves the user-facing documentation for five V2 connectors: Doris, Iceberg, StarRocks, Paimon, and Clickhouse. Both the English (docs/en) and Chinese (docs/zh) sources and sinks are updated.

Highlights

Doris

• Fix an unclosed markdown code fence before the ### Case-Sensitive Configuration heading in the sink docs (both en/zh) that was swallowing the next heading inside the previous code block.
• Correct the Required flags for schema_save_mode and data_save_mode in the sink options table — both are factory-required, both were marked no.
• Change the save_mode_create_template placeholder from the deprecated ${table_name} to the canonical ${table}, and document table_name as the deprecated alias.
• Add the missing default-database and multi_table_sink_replica options to the sink options table.
• Expand the description of needs_unsupported_type_casting to mention its catalog-pre-processing effect.
• Add database, table, doris.batch.size, and case_sensitive to the source Base configuration table.
• Add MAP to the source Data Type Mapping.
• Add a CDC note explaining Doris source is batch-only.
• Port the entire Multiple table section and add three new e2e-backed examples to the zh sink: CDC + Zeta timer flush, Custom SQL + save_mode_create_template, and Direct To BE + 2PC + table.identifier.

Iceberg

• Add [x] [support multiple table read] to the source feature list.
• Add [x] [exactly-once] to the sink feature list (sink relies on IcebergFilesCommitter two-phase commit).
• Correct catalog_name / namespace defaults to default in both source and sink.
• Document table / table_list as mutually exclusive.
• Enumerate the per-table keys accepted inside table_list (namespace, query, snapshot/stream options).
• Clarify that any table alias works in query (the literal table requirement was misleading).
• Fix the HOCON : → = syntax bug in the AWS S3 Tables REST catalog example in both en and zh.
• Normalise zh Required-column language (yes/no → 是/否).
• Add new e2e-backed examples: Query Filter on a Single Table and Query Filter per Table in table_list.

StarRocks

• Fix the zh sink capability matrix: [ ] [定时刷新] → [x] [定时刷新] (Zeta supports timer flush and the rest of the zh page documents it).
• Correct batch_max_rows / batch_max_bytes types: batch_max_rows is int (not long), batch_max_bytes is long (not int), in both en and zh.
• Document table / table_list as mutually exclusive in source docs (en/zh).

Paimon

• Add [x] [parallelism] (Paimon source supports parallel source splits) and [x] [support multiple table read] to the source feature list.
• Flip the source database row from Required 是/Yes to 否/No (factory treats it as optional).
• Fix paimon.hadoop.conf type label from [string] to [Map].
• Rename the duplicate ## Changelog heading to ## Reading Paimon Changelog (source) and ## Changelog Producer (sink) so it no longer collides with the <ChangeLog /> footer.
• Add the missing multi_table_sink_replica option to the sink docs (en/zh).

Clickhouse

• Add language tags (hocon) and SQL tags (sql) to the multi-table sink example in both en and zh.
• Add a parallelism capability row to the en sink feature list (the connector uses the standard parallel sink writer).
• Add a default database option to the source options table for the case where neither table_list nor sql specifies one.

Files touched

• docs/en/connectors/source/{Clickhouse,Doris,Iceberg,Paimon,StarRocks}.md
• docs/en/connectors/sink/{Clickhouse,Doris,Iceberg,Paimon,StarRocks}.md
• docs/zh/connectors/source/{Clickhouse,Doris,Iceberg,Paimon,StarRocks}.md
• docs/zh/connectors/sink/{Clickhouse,Doris,Iceberg,Paimon,StarRocks}.md

20 files changed, 648 insertions(+), 177 deletions(-)