Eventual-Inc
Daft
Blog
Docs
Changelog
Blog
Docs
Changelog
Overview
Branches
Benchmarks
Runs
Performance History
Latest Results
Merge remote-tracking branch 'origin/main' into feat/issue-components-labels
BABTUNA:feat/issue-components-labels
9 hours ago
extend file ref
chenghuichen:file-extend
9 hours ago
fix(ci): avoid stripping labels on issue edits Run component auto-labeling only on issue open events and skip management when Component(s) section is absent to prevent removing manually-applied labels on free-form/template-less issues.
BABTUNA:feat/issue-components-labels
9 hours ago
extend file ref
chenghuichen:file-extend
10 hours ago
paimon read fallback when has deletion vectors or blob file
gavin9402:fix_paimon_deletion_vector_read
11 hours ago
feat(list): add list_flatten expression (#6812) ## Changes Made <!-- Describe what changes were made and why. Include implementation details if necessary. --> This PR implements the `list_flatten` feature to allow reducing the nesting level of list columns by only one level. `list_flatten` maintains the original row count and only flattens the inner list structure. I've implemented the kernel to handle standard `List` and `FixedSizeList` types. The implementation carefully handles edge cases like outer nulls remain null and empty lists are preserved, avoiding the row-generation behavior of `explode`. I also integrated the function into the `Expressions API`, `Series namespace`, and `SQL interface`. I added tests at the `RecordBatch` level to verify the flattening logic, as well as `SQL` and `Series-level` tests to confirm the API surface works as expected. ## Related Issues Closes #6622 <!-- Link to related GitHub issues, e.g., "Closes #123" -->
main
11 hours ago
write_iceberg: clarify error message and retry-narrowness comment Two prose tightening fixes from a final pre-PR code review: * The single-query-id invariant error message previously implied "concurrent writers" or "store corruption" as the only causes. Add "store reused across destinations" and "pre-feature legacy entries with empty query_id" to the list, and point at the resolution (fresh path or clear the store). Less alarming for the common case; more actionable. * Add a comment on the retry loop's `except CommitFailedException` explaining the narrowness is intentional — only optimistic-concurrency conflicts retry; other transient catalog errors (REST 5xx, network blips, auth) propagate to the caller for application-level retry. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
rohit/feature/iceberg-idempotent-snapshot-commits
11 hours ago
feat(native): emit TaskStart events from native executor
chris/task-start-event
11 hours ago
Latest Branches
CodSpeed Performance Gauge
-1%
ci(issues): auto-label issues from component selections
#6820
9 hours ago
bf21002
BABTUNA:feat/issue-components-labels
CodSpeed Performance Gauge
0%
feat(file): add byte-range read support to FileReference
#6842
9 hours ago
f12d06e
chenghuichen:file-extend
CodSpeed Performance Gauge
0%
fix: paimon read fallback when has deletion vectors or blob file
#6839
12 hours ago
b6359e2
gavin9402:fix_paimon_deletion_vector_read
© 2026 CodSpeed Technology
Home
Terms
Privacy
Docs