boluor opened a new pull request, #3794: URL: https://github.com/apache/doris-website/pull/3794
## Summary Doc page (4.x): \`aggregate-functions/corr-welford.md\` (EN + ZH). Two problems on this page; this PR fixes both for both sides. **1. Non-deterministic result tables.** The shared examples on both sides used hash-partitioned tables without ORDER BY: \`\`\`sql select * from test_corr; select id, corr_welford(k1,k2) from test_corr group by id; \`\`\` The documented row orders did not match what the cluster actually produces on a single-node Apache Doris 4.1.1 cluster (EN's documented \`select *\` order was \`(1,20,22),(1,10,20)\` but the cluster returns \`(1,10,20),(1,20,22)\`; the \`group by\` was even more divergent — EN had \`2,4,3,1\`, ZH had \`1,2,3,4\`, cluster gave \`3,1,2,4\`). Added \`ORDER BY\` to both queries so users running the snippet see the same rows as the doc. **2. Each side was missing one example the other already had.** - EN gained ZH's \"no valid data → NULL\" example (\`where id = 999\`) and the matching \"If the group has no valid data, returns NULL\" return-value bullet. - ZH gained EN's \`select * from test_corr\` data-sample step. After this change both pages have the same 4-example structure: setup → show data → group-by → no-data. ## Verification Every SQL example was run against a single-node Apache Doris 4.1.1 cluster — the result blocks in this PR are the verbatim mysql client output. ## Test plan - [x] Run the full setup + 4 examples on the cluster — all match. - [x] Compare EN and ZH structurally — same 4 examples in the same order. - [x] No prose changes beyond the one-line intro added to the no-data example. 🤖 Generated with [Claude Code](https://claude.com/claude-code) -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
