If else condition inside df.loc pandas

user_2727167 0
Viewed 165

I have following dataframe:

                                  time      u10  ...         tsn        tp
longitude latitude                               ...                      
20.0      45.0     2014-01-01 00:00:00 2.595551  ...  272.453827  0.000014
          45.0     2014-01-01 01:00:00 2.615493  ...  273.159973  0.000000
          45.0     2014-01-01 02:00:00 2.587403  ...  273.122192  0.000000
          45.0     2014-01-01 03:00:00 2.528865  ...  273.050903  0.000000
          45.0     2014-01-01 04:00:00 2.556740  ...  272.772491  0.000000

I want to subtract neighboring records of column u10 for all values of column time,except one value of column time ( where time ends with 00:00:00 )

I need following output:

                                  time      u10  ...         tsn        tp
longitude latitude                               ...                      
20.0      45.0     2014-01-01 00:00:00 2.595551  ...  272.453827  0.000014
          45.0     2014-01-01 01:00:00 0.019942  ...  273.159973  0.000000
          45.0     2014-01-01 02:00:00 -0.02809  ...  273.122192  0.000000
          45.0     2014-01-01 03:00:00 -0.058538  ...  273.050903  0.000000

I can do df.loc combining with df['time'].shift()-df['time'] but that will work for all record. How can I make this work with desired output? P.S. I am looking for vectorized solution.

pandas python
1 Answers

Use np.where. Where midnight occurs, keep u10 values, where it doesnt find the consecutive differences.

df['u10']=np.where(df['time'].dt.time==time(0,0,0), df['u10'], df['u10'].diff())

                              time       u10
longitude latitude                              
20.0      45.0     2014-01-01 00:00:00  2.595551
          45.0     2014-01-01 01:00:00  0.019942
          45.0     2014-01-01 02:00:00 -0.028090
          45.0     2014-01-01 03:00:00 -0.058538
          45.0     2014-01-01 04:00:00  0.027875
user_89869750
Related