zhangshenghang opened a new pull request, #12553:
URL: https://github.com/apache/seatunnel/pull/12553
## What documentation issues were found
A systematic comparison of the option tables in every
`docs/en/connectors/{source,sink}` page against its `docs/zh` counterpart (207
doc pairs) found 12 real documentation issues: 11 option tables out of sync
between the English and Chinese docs, and one Chinese doc missing entirely.
Every option name, type, required flag, and default value added by this PR was
verified against the connector source code before being written.
### Issues fixed
**Chinese docs missing options that exist in English docs and in code**
- `zh sink HdfsFile`: table missing 7 sink options: `sheet_max_rows`,
`xml_root_tag`, `xml_row_tag`, `xml_use_attr_format`,
`parquet_avro_write_timestamp_as_int96`, `parquet_avro_write_fixed_as_int96`,
`encoding` (all confirmed in `FileBaseSinkOptions.java`)
- `zh sink S3File`: table missing `filename_extension` and `sheet_max_rows`;
detailed-options section missing `sheet_max_rows`
- `zh sink ObsFile`: table missing `sheet_max_rows`
- `zh sink OssFile`: table missing `filename_extension`
- `zh sink SftpFile`: table missing `filename_extension`
- `zh sink StarRocks`: table missing `table_options` (confirmed in
`SinkConnectorCommonOptions.TABLE_OPTIONS`; the zh detailed-options section
already existed but the table row did not)
- `zh source BosFile` / `zh sink BosFile`: tables missing the
`common-options` row
- `zh source PostgreSQL`: table missing the `common-options` row, plus a
stray lone `|` line rendering as a broken empty table row
**English docs missing options that exist in Chinese docs and in code**
- `en sink Greenplum`: table missing 5 JDBC sink options present in the zh
page: `primary_keys`, `connection_check_timeout_sec` (default 30),
`max_commit_attempts` (default 3), `transaction_timeout_sec` (default -1),
`enable_upsert` (default true) — defaults verified in `JdbcSinkOptions.java` /
`JdbcCommonOptions.java`
- `en source OssJindoFile`: table missing `filename_extension` (the `###
filename_extension` section already existed; only the table row was missing)
**Missing Chinese page**
- `zh source Splunk`: no Chinese page existed at all, so the zh site sidebar
(auto-generated from `docs/zh/connectors/source`) had no Splunk entry. Added a
full Chinese translation of the English page, plus the missing
`docs/zh/connectors/changelog/connector-http-splunk.md` required by the page's
changelog import.
## What was changed
- Added the missing option-table rows listed above, with Chinese
descriptions translated from the (code-verified) English pages, placed in the
same relative order as the English tables
- Fixed the malformed trailing table row in `zh source PostgreSQL`
- Added the missing Chinese Splunk source page and its changelog file
## Areas / files updated
13 files: `docs/en/connectors/sink/Greenplum.md`,
`docs/en/connectors/source/OssJindoFile.md`,
`docs/zh/connectors/sink/{BosFile,HdfsFile,ObsFile,OssFile,S3File,SftpFile,StarRocks}.md`,
`docs/zh/connectors/source/{BosFile,PostgreSQL,Splunk}.md`,
`docs/zh/connectors/changelog/connector-http-splunk.md`
## Were both English and Chinese docs checked?
Yes. The comparison script covered all 207 en/zh connector doc pairs (source
+ sink + transform). After the fix, the option-table diff between English and
Chinese docs is zero across all pairs. Two issues were in the English docs
(Greenplum sink, OssJindoFile source); the rest were in the Chinese docs.
## Duplicate PR check (last 7 days)
Checked open + merged PRs from 2026-09-22 onward with `gh pr list --search
"docs"`. Related docs PRs reviewed for overlap:
- #12459 (open): Jdbc-family connectors, Zeta engine docs — no file overlap
- #12499 (merged): file connector docs — this PR only fixes remaining
zh-table gaps that survived in current `dev` (verified against latest
upstream/dev, which includes #12499); no content duplication
- #12514 (open): CosFile source table format — no overlap (CosFile not
touched)
- #12447 / #12476 / #12487 / #12505 / #12452 / #12427 (merged):
Clickhouse/transform/Doris/DuckDB/JDBC-links areas — no file overlap
- No docs PR from today overlaps this scope
## Verification
- Option-table comparison script across all 207 en/zh connector pairs: 0
remaining differences
- Markdown table column-count validation on all 12 edited pages: 0 problems
- Relative-link check across all connector docs (en + zh): 0 broken links
- External links: no new external links introduced (CI `markdown-link-check`
scope unaffected)
- Full site build not run locally (requires the docusaurus toolchain); the
changed files are pure markdown table rows/sections following the existing page
conventions
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
