|
Repo Added
|
Files
47
|
Badge
README BADGES
|
push
github
fix: preserve zero-column metadata rows (DAL-192) (#262) Preserve row counts when normalizing Arrow metadata on zero-column inputs. Use RecordBatch metadata replacement before recombining these inputs, while retaining the existing buffer-sharing path for tables with columns. Cover Table, RecordBatch and native Batch inputs, empty and populated inputs, sync/async collection, iterable streams, caller metadata and resource cleanup. Keep the stream regression below the existing complexity threshold and document the corrected behavior. Validated on head 760f071f6 against main 6a09e9233: 52 focused tests pass; the full Python suite passes 1529 tests with 2 optional-feature skips and 95.25% coverage. Final specialist review approves the change, and all 53 CI checks pass. Closes DAL-192
9518 of 10098 relevant lines covered (94.26%)
0.94 hits per line
| Coverage | ∆ | File | Lines | Relevant | Covered | Missed | Hits/Line |
|---|