Skip to content
Draft
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
78 changes: 78 additions & 0 deletions python/tests/test_merge.py
Original file line number Diff line number Diff line change
Expand Up @@ -2297,3 +2297,81 @@ def test_merge(tmp_path: pathlib.Path):

new_df = pl.read_delta(str(tmp_path))
assert_frame_equal(df, new_df, check_row_order=False)


@pytest.mark.pyarrow
@pytest.mark.pandas
def test_merge_file_pruning_regression_3636(tmp_path: pathlib.Path):
"""
https://github.com/delta-io/delta-rs/issues/3636
"""
import pyarrow as pa
import pandas as pd
import numpy as np
import deltalake
from deltalake import DeltaTable

# create table once
dt = DeltaTable.create(
tmp_path,
pa.schema(
[
pa.field("ts", pa.timestamp("us"), nullable=False),
pa.field("direction", pa.string(), nullable=False), # S, W, N, E
pa.field("distance", pa.float64(), nullable=False),
pa.field("updated_on", pa.int64(), nullable=False), # timestamp in ms
pa.field("date", pa.string(), nullable=False),
]
),
partition_by=["date"],
mode="ignore",
)

def mock_data(size: int):
tss = pd.date_range(start="2025-01-01", freq="2s", periods=size // 4)

init_upd_df = {
"ts": [*tss] * 4,
"direction": [i for i in ("S", "W", "N", "E") for _ in range(len(tss))],
"distance": np.random.randint(0, 10000, size=len(tss) * 4),
"updated_on": list(
(tss.floor("D") + pd.Timedelta(hours=25)).astype(int) // 1000000
)
* 4,
"date": list(tss.strftime("%Y-%m-%d")) * 4,
}
return pd.DataFrame(init_upd_df)

# Let's prefill the table large enough to spread the results across
# multiple files
init_df = mock_data(500_000)

dt = DeltaTable(tmp_path)

dt.merge(
init_df,
predicate="s.date = t.date and s.ts = t.ts and s.direction = t.direction",
source_alias="s",
target_alias="t",
).when_not_matched_insert_all().execute()

second_upd = mock_data(40)

dt = DeltaTable(tmp_path)

stats = (
dt.merge(
second_upd,
predicate="s.date = t.date and s.ts = t.ts and s.direction = t.direction",
source_alias="s",
target_alias="t",
)
.when_not_matched_insert_all()
.execute()
)

files_scanned = stats["num_target_files_scanned"]

assert files_scanned <= 1, (
f"The number of target files scanned was too large! {files_scanned}"
)
Loading