pandas/row-loop
| Rule | row-loop |
|---|---|
| Pack | pandas |
| Severity | warning |
| Fix | none, edit by hand |
| Library versions | every version |
iterrows() builds a Series for every row, and a loop that only computes from row values runs Python once for every row. Compute on whole columns instead.
total = sum(row.load * row.weight for row in df.itertuples())total = (df["load"] * df["weight"]).sum()What the rule reports
Section titled “What the rule reports”The rule reports a loop over iterrows(), a loop over itertuples() or to_dict("records") that only computes from the row values, and for i in range(len(df)) with df.iloc[i]. It does not run on test files.
The rule leaves alone a loop over itertuples() or to_dict("records") that hands the row or its values to a function, as in orders = [Order(id=row["id"], total=row["total"]) for row in df.to_dict("records")], because the loop turns rows into Python objects. A cast such as int(row["a"]) and a list method such as out.append(...) don't count as a function here. The rule reports iterrows() in every loop, because iterrows() builds a Series for each row. Use itertuples() or to_dict("records") when each row has to become an object.
How to fix it
Section titled “How to fix it”xpd has no automatic fix, so make the change by hand.
The replacement sums numeric columns without missing values. When you replace a Python reduction, check the missing values and the floating point rounding. Keep a loop when its work can't be written on columns.
Message
Section titled “Message”xpd prints this message with the issue. A name that starts with $ stands for code from your file.
iterrows() builds a Series for every row. Compute on whole columns instead, such as df["a"] * df["w"]. When each row has to become a Python object, loop over itertuples() or to_dict("records").Browse the other rules of the pandas pack.