numpy/sort-then-select
| Rule | sort-then-select |
|---|---|
| Pack | NumPy |
| Severity | warning |
| Fix | unsafe, with xpd check --fix --unsafe-fixes |
| Library versions | every version |
Sorting every value to read the smallest or largest few was up to 39 times slower. Use np.min, np.max, np.partition or np.argpartition.
top = np.sort(scores)[-5:]top = np.sort(np.partition(scores, -5)[-5:])What the rule reports
Section titled “What the rule reports”The rule reports np.sort(x)[0], np.sort(x)[-1], np.sort(x)[:5], np.sort(x)[-5:], np.sort(x)[::-1][:5], x[np.argsort(x)], np.argsort(x)[0], np.argsort(x)[-1], np.argsort(x)[:5], np.argsort(x)[-5:] and np.argsort(np.argsort(x)), with a literal count under 100.
How to fix it
Section titled “How to fix it”xpd check --fix --unsafe-fixes applies the fixes of this rule. Check the result, because an unsafe fix can change what the code does.
The unsafe fix writes np.min(x), np.max(x), np.sort(np.partition(x, 5)[:5]), np.sort(np.partition(x, -5)[-5:]), np.sort(x), np.argmin(x), np.argmax(x). The positions of the k smallest or largest and the rank by a double argsort have no fix.
Message
Section titled “Message”xpd prints this message with the issue. A name that starts with $ stands for code from your file.
This sorts every value to read the smallest one, which ran 39 times slower on 1,000,000 values. Write $NP.min($X).Browse the other rules of the NumPy pack.