Skip to content

pandas/index-time-mask

Rule index-time-mask
Pack pandas
Severity warning
Fix unsafe, with xpd check --fix --unsafe-fixes
Library versions every version

Selecting rows by index.time or index.hour builds a Python time or an hour for every row. Use at_time or between_time instead.

noon = ts[ts.index.time == time(12)]
working = ts[(ts.index.hour >= 8) & (ts.index.hour < 16)]
noon = ts.at_time(time(12))
working = ts.between_time("8:00", "16:00", inclusive="left")

The rule reports ts[ts.index.time == time(12)] and ts[(ts.index.hour >= 8) & (ts.index.hour < 16)], which build a Python time or an hour for every row. The table shows the times on 2,000,000 rows.

Mask Time Replacement Time
ts.index.time == time(12) 241 ms at_time 25 ms
the two hour masks 51 ms between_time 28 ms

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 safe fix writes ts.at_time(time(12)). The unsafe fix writes ts.between_time("8:00", "16:00", inclusive="left").

The at_time fix applies to a time object, such as time(12). The rule leaves a comparison with text alone, because text never equals a time. The between_time fix applies to whole hours written as numbers. It is unsafe when the first hour is later than the second. Then the masks select nothing, and between_time selects the hours across midnight.

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

index.time builds a Python time object for every row. Write $X.at_time($T).

Browse the other rules of the pandas pack.