Eventual-Inc/Daft

Daft

Performance History

Latest Results

fix: don't swallow unrelated create failures in create_table_if_not_exists create_table_if_not_exists caught every ValueError from create_table and treated it as "table already exists", which could mask unrelated creation failures (e.g. connection or permission errors raised by Python-backed catalogs like Postgres) behind a misleading not-found error or a false success. Only recover from errors whose message contains "already exists"; other failures propagate unchanged. Adds a regression test.

Lucas61000:issue-7310

9 minutes ago

fix: eliminate TOCTOU race in CREATE TABLE IF NOT EXISTS Fix three TOCTOU (time-of-check-time-of-use) windows where concurrent callers of CREATE TABLE IF NOT EXISTS could see 'already exists' errors: 1. MemoryCatalog::create_table (memory.rs): replace separate read-lock check + write-lock insert with a single write lock for the entire check-then-insert, so two concurrent creators cannot both pass the existence check. The table is built before taking the lock because MemoryTable::new registers the partition set with the (Python) runner and can release the GIL; holding the lock across that callback can deadlock with a concurrent creator. 2. SQL executor (exec.rs): try create_table first, catch CatalogError::ObjectAlreadyExists when IF NOT EXISTS is set. Also handles Python-backed catalogs where the error wraps as CatalogError::PythonError. 3. Python Catalog API (__init__.py): try create_table first, catch ValueError (DaftCoreException) and return the existing table. References Spark's CreateTableExec which catches TableAlreadyExistsException after create for the same reason. Closes #7310

Lucas61000:issue-7310

2 hours ago

fix: simplify actor UDF repartition handling

everettVT/fix-main-failures

11 hours ago

Merge branch 'Eventual-Inc:main' into main

ARDA7787:main

12 hours ago

fix(udf): preserve field names when unnesting with_column

dubin555:contrib/issue-5448

15 hours ago

fix(utf8): return empty strings for empty substr results

dubin555:contrib/issue-3700

15 hours ago

docs(lerobot): drop dataset-specific names from comments and docs Keep v2/v3 layout docs general instead of calling out particular Hub datasets.

XuQianJin-Stars:feat/lerobot-v2

2 days ago

feat(utf8): add split_part string function (#7371) ## Changes Made Daft has a split function that returns a list of parts, but no direct way to split a string on a delimiter and select the nth part. This PR adds split_part, which does exactly that with the same name and behavior as the function in PostgreSQL, DuckDB, Snowflake and Spark: split_part(input, delim, part) returns the requested part, 1-based. Negative parts count from the end, out of range parts return an empty string, part 0 raises an error, and an empty delimiter means the string is not split. The implementation follows the same structure as substring_index and translate from #7070: - `src/daft-functions-utf8/src/split_part.rs`: Rust kernel with unit tests - `src/daft-functions-utf8/src/lib.rs`: registration, which also enables the SQL function - `daft/functions/str.py`: Python API function with doctests - `daft/functions/__init__.py`: export - `daft/expressions/expressions.py`: Expression method - `tests/expressions/test_utf8.py`: tests covering the DataFrame, SQL, RecordBatch and Series paths Naming: #7350 has an open discussion on whether a different name or an argument on split would be clearer. I kept split_part since it is the common name across engines, and I am happy to rename or adjust. Test evidence: ``` cargo test -p daft-functions-utf8 split_part test result: ok. 11 passed; 0 failed; 0 ignored DAFT_RUNNER=native pytest tests/expressions/test_utf8.py 21 passed pytest --doctest-modules daft/functions/str.py::daft.functions.str.split_part 1 passed ``` All pre-commit hooks pass, including mypy, ruff, rustfmt and clippy. AI usage: I used Claude while working on this, mostly to help me follow the existing patterns and keep the code consistent with the other string functions. I read and tested everything myself before opening this PR. ## Related Issues Fixes #7350

main

2 days ago

Latest Branches

+1%

fix: eliminate TOCTOU race in CREATE TABLE IF NOT EXISTS#7387

30 minutes ago

3b67604

Lucas61000:issue-7310

fix: clean up unscheduled UDF actors#7346

11 hours ago

886f652

everettVT/fix-main-failures

feat(dataframe): add query profiling#7297

12 hours ago

c93bfe9

ARDA7787:main

Home Terms Privacy Docs