Skip to content

Commit a07e2b9

Browse files
authored
Remove stale files and prune CLAUDE.md of nonexistent code references (apache#115)
- Delete claude-progress.txt, QUICK-START.md, task_list.json, orc-predicate-pushdown.allium - Remove CLAUDE.md references to OrcFileFragment, OrcSchemaManifest, BuildOrcSchemaManifest, GetOrcColumnIndex, StripeIterationLogic (none exist) - Remove "Python-Level Gaps" section referencing nonexistent predicate pushdown types - Fix test file line counts to match actual source
1 parent b0d2f91 commit a07e2b9

5 files changed

Lines changed: 2 additions & 3071 deletions

File tree

CLAUDE.md

Lines changed: 2 additions & 19 deletions
Original file line numberDiff line numberDiff line change
@@ -96,8 +96,8 @@ ctest --label-regex unittest --output-on-failure --parallel <nproc> --repeat unt
9696

9797
| Binary | Source File | What It Tests |
9898
|--------|------------|---------------|
99-
| `arrow-orc-adapter-test` | `cpp/src/arrow/adapters/orc/adapter_test.cc` (~1,259 lines) | Low-level adapter: type conversion, read/write roundtrip, column statistics extraction |
100-
| `arrow-dataset-file-orc-test` | `cpp/src/arrow/dataset/file_orc_test.cc` (~331 lines) | Dataset layer: format inspection, scanning, schema manifest building, column index resolution |
99+
| `arrow-orc-adapter-test` | `cpp/src/arrow/adapters/orc/adapter_test.cc` (~1,182 lines) | Low-level adapter: type conversion, read/write roundtrip, column statistics extraction |
100+
| `arrow-dataset-file-orc-test` | `cpp/src/arrow/dataset/file_orc_test.cc` (~96 lines) | Dataset layer: format inspection, scanning |
101101

102102
### C++ Test Patterns
103103

@@ -107,19 +107,6 @@ Tests use Google Test with Arrow's fixture mixin framework:
107107
- **`TestOrcFileFormat`** (inherits `FileFormatFixtureMixin<OrcFormatHelper>`) — basic format tests: `Inspect`, `IsSupported`, `CountRows`, `FragmentEquals`
108108
- **`TestOrcFileFormatScan`** (inherits `FileFormatScanMixin<OrcFormatHelper>`, parameterized) — scan tests: `ScanRecordBatchReader`, `ScanBatchSize`, `ScanProjected`, `ScanProjectedNested`, `ScanWithPushdownNulls`, etc.
109109

110-
Custom tests for the predicate pushdown work:
111-
- `BuildSchemaManifest_FlatSchema` / `BuildSchemaManifest_NestedSchema` — verifies Arrow-to-ORC column index mapping
112-
- `GetOrcColumnIndex_TopLevelFields` / `GetOrcColumnIndex_NestedFields` — verifies `FieldRef` resolution through the manifest
113-
- `StripeIterationLogic` — placeholder confirming empty stripe list = read-all path
114-
115-
These custom tests forward-declare internal functions for white-box testing:
116-
```cpp
117-
namespace internal {
118-
Result<std::shared_ptr<OrcSchemaManifest>> BuildOrcSchemaManifest(...);
119-
std::optional<int64_t> GetOrcColumnIndex(...);
120-
}
121-
```
122-
123110
### CMake Wiring
124111

125112
In `cpp/src/arrow/dataset/CMakeLists.txt`:
@@ -225,10 +212,6 @@ Located in `python/pyarrow/tests/data/orc/`:
225212
- Corresponding `*.jsn.gz` files with expected values
226213
- Sourced from Apache ORC examples: https://github.com/apache/orc/tree/main/examples
227214

228-
### Python-Level Gaps
229-
230-
The new C++ predicate pushdown types (`OrcFileFragment`, `OrcSchemaManifest`, statistics APIs, etc.) have **no Python bindings yet**. The existing Cython files only wrap the original upstream `ORCFileReader`/`ORCFileWriter`. Exposing predicate pushdown to Python would require new declarations in `_orc.pxd` and wrappers in `_dataset_orc.pyx`.
231-
232215
---
233216

234217
## CI Configuration

QUICK-START.md

Lines changed: 0 additions & 183 deletions
This file was deleted.

claude-progress.txt

Lines changed: 0 additions & 68 deletions
This file was deleted.

0 commit comments

Comments
 (0)