Skip to content

Commit

Permalink
add parquet benchmark
Browse files Browse the repository at this point in the history
  • Loading branch information
samster25 committed Sep 20, 2023
1 parent 0fb351b commit 0b2d52b
Showing 1 changed file with 6 additions and 0 deletions.
6 changes: 6 additions & 0 deletions benchmarking/parquet/conftest.py
Original file line number Diff line number Diff line change
Expand Up @@ -80,6 +80,10 @@ def daft_bulk_read(paths: list[str], columns: list[str] | None = None) -> list[p
return [t.to_arrow() for t in tables]


def daft_into_pyarrow_bulk_read(paths: list[str], columns: list[str] | None = None) -> list[pa.Table]:
return daft.table.read_parquet_into_pyarrow_bulk(paths, columns=columns)


def pyarrow_bulk_read(paths: list[str], columns: list[str] | None = None) -> list[pa.Table]:
return [pyarrow_read(f, columns=columns) for f in paths]

Expand All @@ -91,11 +95,13 @@ def boto_bulk_read(paths: list[str], columns: list[str] | None = None) -> list[p
@pytest.fixture(
params=[
daft_bulk_read,
daft_into_pyarrow_bulk_read,
pyarrow_bulk_read,
boto_bulk_read,
],
ids=[
"daft_bulk_read",
"daft_into_pyarrow_bulk_read",
"pyarrow_bulk_read",
"boto3_bulk_read",
],
Expand Down

0 comments on commit 0b2d52b

Please sign in to comment.