将函数应用于数据框中的多列范围

Question

你能告诉我如何使用像x['D'] :x['A']这样的列范围，而不是将所有列作为参数在下面一一列出来运行吗？

df.loc[:,'F']=df.apply(lambda x: my_function(x['D'],x['C'],x['B'],x['A']), axis=1)

Answer 1

Were you looking for something like this?

df.loc[:,'F']=df.apply(lambda x: my_function(x.loc['D':'A']), axis=1)

If you need the parameters to be a list of columns and not a dataframe:

df.loc[:,'F']=df.apply(lambda x: my_function([x for x in x.loc['D':'A']]), axis=1)

Answer 2

You can try do something like this:

df.loc[:,'F']=df.apply(lambda x: my_function(*df.columns.difference(['F'])), axis=1)

if your columns are in order you can slice like this:

df.loc[:,'F']=df.apply(lambda x: my_function(*df.iloc[:,1:4])), axis=1)

Answer 3

Say you have a dataframe like this

    E   D   C   B   A
______________________

0   8   9   9   6   1
1   8   4   1   4   3
2   0   4   6   9   3

And you want to apply some function func to columns D to A , you can try

def func(x):
    return sum(x[1:])

df.apply(func, axis=1)

output

0    25
1    12
2    22
dtype: int64