CODEWITHSUNDEEP
pythonpandasdataframe
Simon
Simon

Reputation: 10158

Python / pandas - convert conditional hours to minutes

I have some working code to convert hours to minutes (with a condition on the hour value) but I'm currently iterating over the dataframe rows and that doesn't seem very efficient for larger dataframes

I have 2 pandas series (in reality these are columns of a larger dataframe):

hours = pd.Series([0,10,15,20,30])
mins = pd.Series([10,0,0,20,10])

What I want to do is the following:

I'm expecting the output to be the following series (note indices 2 and 4):

0      10.0
1     600.0
2      15.0
3    1220.0
4      40.0

I wrote the following function that achieves what I want:

def convert_time(hours, minutes):
    df_duration = pd.DataFrame({"hours": hours, "minutes": minutes}).astype(float)
    series_converted = pd.Series()

    # Replace nan with zero
    df_duration = df_duration.fillna(0)

    # Convert out of bound hour values, add to minutes
    for i in range(df_duration.shape[0]):
        if df_duration.iloc[i]["hours"] in (15,30,45,60,90):
            cur_hours = 0
            cur_mins = df_duration.iloc[i]["hours"] + df_duration.iloc[i]["minutes"]
        else:
            cur_hours = df_duration.iloc[i]["hours"]
            cur_mins = df_duration.iloc[i]["minutes"]

        series_converted.set_value(i,(cur_hours * 60) + cur_mins)

    return series_converted

Whats the correct way to do this in Pandas without iterating over dataframe rows? I can't seem to figure out how to use pandas masking across 2 columns when a condition is involved

Upvotes: 1

Views: 208

Answers (1)

EdChum
EdChum

Reputation: 394459

You can use where with isin to only perform the conversion where the values are not in the list of values you want, then you can fillna with the normal operation:

In [134]:
(hours.where(~hours.isin([15,30,45,60,90])) * 60 + mins).fillna(hours + mins)

Out[134]:
0      10.0
1     600.0
2      15.0
3    1220.0
4      40.0
dtype: float64

You can see this broken down:

In [135]:   
hours.isin([15,30,45,60,90])

Out[135]:
0    False
1    False
2     True
3    False
4     True
dtype: bool

In [136]:
~hours.isin([15,30,45,60,90])

Out[136]:
0     True
1     True
2    False
3     True
4    False
dtype: bool

In [138]:
hours.where(~hours.isin([15,30,45,60,90]))

Out[138]:
0     0.0
1    10.0
2     NaN
3    20.0
4     NaN
dtype: float64

In [139]:
(hours.where(~hours.isin([15,30,45,60,90])) * 60 + mins)

Out[139]:
0      10.0
1     600.0
2       NaN
3    1220.0
4       NaN
dtype: float64

Upvotes: 2

Related Questions