QT/superset2
1
0
Fork 0
mirror of https://github.com/apache/superset.git synced 2026-05-21 15:55:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(charts): use separator-aware matching for time shift column grouping

Browse Source 
The `endswith(ts)` check in `get_column_groups` matched columns
incorrectly when time shifts shared a numeric suffix — e.g.
"metric__22 weeks ago".endswith("2 weeks ago") returned True,
causing columns to be misclassified. Use `endswith(TIME_COMPARISON + ts)`
to include the "__" separator in the match.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Maxime Beauchemin
2026-04-13 17:19:10 +00:00
committed by Sophie
parent 4aa4415d8f
commit 6c00ee2eec
2 changed files with 35 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
superset/utils/pandas_postprocessing/contribution.py
View File

@@ -24,7 +24,7 @@ from flask_babel import gettext as _
from pandas import DataFrame, MultiIndex
from superset.exceptions import InvalidPostProcessingError
from superset.utils.core import PostProcessingContributionOrientation
from superset.utils.core import PostProcessingContributionOrientation, TIME_COMPARISON
from superset.utils.pandas_postprocessing.utils import validate_column_args
@@ -130,7 +130,7 @@ def get_column_groups(
time_shift = None
if time_shifts and isinstance(col_0, str):
for ts in time_shifts:
if col_0.endswith(ts):
if col_0.endswith(TIME_COMPARISON + ts):
time_shift = ts
break
if time_shift is not None:

33
tests/unit_tests/pandas_postprocessing/test_contribution.py
View File

@@ -124,3 +124,36 @@ def test_contribution_with_time_shift_columns():
assert_array_equal(processed_df["a__1 week ago"].tolist(), [0.5, 0.5])
assert_array_equal(processed_df["b__1 week ago"].tolist(), [0.25, 0.25])
assert_array_equal(processed_df["c__1 week ago"].tolist(), [0.25, 0.25])
def test_contribution_with_numeric_prefix_time_shifts():
"""Time shifts like '2 weeks ago' and '22 weeks ago' share a numeric suffix;
columns must be grouped by their exact offset, not by suffix matching."""
df = DataFrame(
{
DTTM_ALIAS: [
datetime(2020, 7, 16, 14, 49),
datetime(2020, 7, 16, 14, 50),
],
"a": [3, 6],
"b": [6, 3],
"a__2 weeks ago": [1, 1],
"b__2 weeks ago": [1, 1],
"a__22 weeks ago": [2, 4],
"b__22 weeks ago": [4, 2],
}
)
processed_df = contribution(
df,
orientation=PostProcessingContributionOrientation.ROW,
time_shifts=["2 weeks ago", "22 weeks ago"],
)
# Non-time-shift columns: a=3,b=6 -> a=1/3, b=2/3; a=6,b=3 -> a=2/3, b=1/3
assert_array_equal(processed_df["a"].tolist(), [1 / 3, 2 / 3])
assert_array_equal(processed_df["b"].tolist(), [2 / 3, 1 / 3])
# "2 weeks ago" group: a=1,b=1 -> 0.5,0.5 each row
assert_array_equal(processed_df["a__2 weeks ago"].tolist(), [0.5, 0.5])
assert_array_equal(processed_df["b__2 weeks ago"].tolist(), [0.5, 0.5])
# "22 weeks ago" group: a=2,b=4 -> 1/3,2/3; a=4,b=2 -> 2/3,1/3
assert_array_equal(processed_df["a__22 weeks ago"].tolist(), [1 / 3, 2 / 3])
assert_array_equal(processed_df["b__22 weeks ago"].tolist(), [2 / 3, 1 / 3])