numpy/mask-per-label
| Rule | mask-per-label |
|---|---|
| Pack | NumPy |
| Severity | warning |
| Fix | none, edit by hand |
| Library versions | every version |
Comparing an array with one label per pass makes one pass for every label. Compare with all labels at once, which gives one row per label.
cleared = np.array([np.where(lot_of == lot, weights, 0.0) for lot in range(lots)])cleared = (lot_of == np.arange(lots)[:, None]) * weightsWhat the rule reports
Section titled “What the rule reports”The rule reports np.array([... a == k ... for k in ...]), and a for loop over labels that
selects with values[labels == k].
How to fix it
Section titled “How to fix it”xpd has no automatic fix, so make the change by hand.
The comprehension compares the same array with one label per pass. Broadcasting compares it
with all labels at once and gives one row per label, and on NumPy 2.5.5 the two forms were
equal. For the loop, pd.DataFrame(values).groupby(labels).median() reduces every group in one
pass. When the body fits a model or solves a problem for each label, loop over
frame.groupby(labels) so that the frame is split once.
Message
Section titled “Message”xpd prints this message with the issue. A name that starts with $ stands for code from your file.
This compares $ARRAY with one label per pass. Compare it with all labels at once, as in $ARRAY == np.arange(n)[:, None], which gives one row per label.Browse the other rules of the NumPy pack.