Imagine I have two pandas data frame as:
import pandas as pd
df1 = {'y1': [1, 2, 3, 4]}
df2 = {'y2': [3, 1, 2, 6]}
What I want is if a value in y2 is greater than the value in y1, I want to set df2['y2'] to the corresponding df['y1']. When I try selecting the corresponding columns like:
df2[df2['y2'] > df1['y1']]
This is returns True rather than the index. I was hoping to do something like:
df2[df2['y2'] > df1['y1']]['y2'] = df1['y1']
Use numpy.where:
In [233]: import numpy as np
In [234]: df1 = pd.DataFrame({'y1': [1, 2, 3, 4]})
In [236]: df2 = pd.DataFrame({'y2': [3, 1, 2, 6]})
In [242]: df2['y2'] = np.where(df2.y2.gt(df1.y1), df1.y1, df2.y2)
In [243]: df2
Out[243]:
y2
0 1
1 1
2 2
3 4
If same index in both DataFrames:
Use DataFrame.loc:
df2.loc[df2['y2'] > df1['y1'], 'y2'] = df1['y1']
print (df2)
y2
0 1
1 1
2 2
3 4
df2['y2'] = df1['y1'].where(df2['y2'].gt(df1['y1']), df2['y2'])
df2['y2'] = df2['y2'].mask(df2['y2'].gt(df1['y1']), df1['y1'])
print (df2)
y2
0 1
1 1
2 2
3 4
np.minimumMaintain all of existing df2 but with updated column values in 'y2'
df2.assign(y2=np.minimum(df1.y1, df2.y2))
y2
0 1
1 1
2 2
3 4
Or just a new dataframe with one column
pd.DataFrame({'y2': np.minimum(df1.y1, df2.y2)})
y2
0 1
1 1
2 2
3 4
