Python 2.7 Subset a groupby dataframe based on another groupby

Question

I have 2 groupby dataframes (AUS_2016 and df_pitLaps, both grouped by driver ID) which have the same column ('lap'), and i am trying to subset AUS_2016 such that it does not contains values in that lap column in df_pitLaps for each driver.

To put it simply, I want to filter out pitlaps (laps with pit stop) for a particular race for each driver

I get a dataframe of True values grouped by driver ID, but I'm not sure how to proceed next.

AUS_2016:

df_pitLaps:

def clean_laps_no_pitlaps(data):
    """Filters out the pit laps."""
    df_pitLaps = df_pitStops.loc[df_pitStops['raceId'].isin(data['raceId'])]
    df_pitLaps.groupby("driverId")["lap"]
    data = data.groupby("driverId")["lap"]

    nopitlaps = lambda x: (
        [(lap != pitlap) for pitlap, lap in itertools.izip(x, data)])

    no_pitlaps_in_data = pd.DataFrame(data.apply(nopitlaps))

    return no_pitlaps_in_data

Calling the function:

clean_laps_no_pitlaps(AUS_2016)

This give this error:

DeprecationWarning: elementwise != comparison failed; this will raise an error in the future.

and below dataframe. I am not sure how to continue on from here to only filter laps for each driver that are True (not pit laps).

SOLUTION:

I managed to resolve it by using another method which does not need groupby. I "vlooked-up" the laps with pit stops to df, then excluded these rows.

def no_pitlaps(df, df_pitLaps):
    """Returns a dataframe that excludes the pit laps of each driver"""

    data_pitlaps_mapped = pd.merge(df, df_pitLaps[['driverId', 'stop', 'lap']], how='left',
        left_on=['driverId','lap'], right_on=['driverId','lap'])

    return data_pitlaps_mapped.loc[~data_pitlaps_mapped.index.isin(data_pitlaps_mapped.dropna(subset = ['stop']).index)] 

Answer 1

如果您只想排除AUS_2016中出现的df_pitLaps中的所有圈数，只需执行以下操作即可：

AUS_2016[~AUS_2016['lap'].isin(df_pitLaps['lap'].unique())]