Skip to content

pandas/positional-key-on-labels

Rule positional-key-on-labels
Pack pandas
Severity error
Fix unsafe, with xpd check --fix --unsafe-fixes
Library versions every version

On pandas 3, an integer key on a Series is a label, so value_counts()[0] and similar lookups raise KeyError unless 0 is one of the labels. Use .iloc[0] for the value at a position.

top = df["region"].value_counts()[0]
first = df.iloc[0][0]
top = df["region"].value_counts().iloc[0]
first = df.iloc[0, 0]

The rule reports df["k"].value_counts()[0], df.iloc[i][0], df.dtypes[0] and a frame reduction such as df.isna().sum()[0], which raise KeyError, because an integer key on a Series is a label on pandas 3. It also reports a negative key, as in mean()[-1].

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 .iloc[0].

The fix is unsafe, because the old form still works on labels that are whole numbers. On such labels value_counts()[0] is the count of the value 0. The same holds for df.dtypes[0] and a frame reduction such as df.sum()[0], which get .iloc[0] too.

xpd prints this message with the issue. A name that starts with $ stands for code from your file.

An integer key on a Series is a label on pandas 3, and value_counts labels its rows with the values, so this raises KeyError unless $N is one of the values. For the count at that position write .iloc[$N].

Browse the other rules of the pandas pack.