pandas/astype-then-to-numpy
| Rule | astype-then-to-numpy |
|---|---|
| Pack | pandas |
| Severity | warning |
| Fix | unsafe, with xpd check --fix --unsafe-fixes |
| Library versions | every version |
astype builds a whole Series only for to_numpy to take its values. Pass the dtype to to_numpy instead.
limit = rows["Economic Maximum"].astype(float).to_numpy()limit = rows["Economic Maximum"].to_numpy(dtype=float)What the rule reports
Section titled “What the rule reports”The rule reports s.astype(float).to_numpy().
It leaves a text dtype alone, such as s.astype(str).to_numpy(), because the two forms differ for it. On pandas 3.0.5 astype(str).to_numpy() gave an object array and to_numpy(dtype=str) a fixed-width one, and a datetime column gave "2025-01-01" with the first and "2025-01-01T00:00:00.000000" with the second.
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 s.to_numpy(dtype=float), because a missing value in a nullable column can convert in another way.
xpd applies the replacement with --fix --unsafe-fixes.
astype builds a second Series with the new dtype only for to_numpy to take its values.
to_numpy takes the dtype itself. On pandas 3.1.1 both forms gave an equal float64 array for an
int column.
Message
Section titled “Message”xpd prints this message with the issue. A name that starts with $ stands for code from your file.
astype builds a Series only to take its values. Pass the dtype to to_numpy, as in $X.to_numpy(dtype=$T).Browse the other rules of the pandas pack.