On Sunday, January 3, 2021 at 8:17:16 PM UTC+2, Codeliner wrote: > On Sunday, January 3, 2021 at 7:08:49 PM UTC+2, Jason Friedman wrote: > > > > > > import numpy as np > > > import pandas as pd > > > from numpy.random import randn > > > df=pd.DataFrame(randn(5,4),['A','B','C','D','E'],['W','X','Y','Z']) > > > > > > W X Y Z > > > A -0.183141 -0.398652 0.909746 0.332105 > > > B -0.587611 -2.046930 1.446886 0.167606 > > > C 1.142661 -0.861617 -0.180631 1.650463 > > > D 1.174805 -0.957653 1.854577 0.335818 > > > E -0.680611 -1.051793 1.448004 -0.490869 > > > > > > is there a way to create a column S - which will copy column column Y > > > values UP- if values in column Y are above 1 - otherwise return new value > > > above zero?.I made this manually: > > > > > > S: > > > > > > A 1.446886 > > > B 1.446886 > > > C 1.854577 > > > D 1.854577 > > > E 1.448004 > > > > > Here's one solution. No consideration to performance. > > import numpy as np > > import pandas as pd > > from numpy.random import randn > > df=pd.DataFrame(randn(5,4),['A','B','C','D','E'],['W','X','Y','Z']) > > print(df) > > > > y_series = df["Y"] > > for i in range(len(y_series)): > > if i == len(y_series) - 1: > > # Last one, nothing to copy > > break > > if y_series[i+1] > 1: > > y_series[i] = y_series[i+1] > > > > df["Y"] = y_series > > print(df) > Thank you Jason for this lovely for loop - is there a way to make this with > pandas series or numpy arrays? for maximum speed?
can something done along these lines? df['run2'] = df['b'].apply(lambda x: df['b'].shift(1) if x > 1 else df['b']) -- https://mail.python.org/mailman/listinfo/python-list
