Pandas: passing multiple column names as arguments to a function with apply

Question

I just asked the following question

Pandas: how can I pass a column name to a function that can then be used in 'apply'?

To which I received a great answer. However, there is an extension to this question that I overlooked and also am curious about.

I have a function:

def generate_confusion_matrix(row):
val=0
if (row['biopsy_bin']==0) & (row['pioped_logit_category'] == 0):
    val = 0   
if (row['biopsy_bin']==1) & (row['pioped_logit_category'] == 1):
    val = 1 
if (row['biopsy_bin']==0) & (row['pioped_logit_category'] == 1):
    val = 2
if (row['biopsy_bin']==1) & (row['pioped_logit_category'] == 0):
    val = 3
if row['pioped_logit_category'] == 2:
    val = 4
return val

I wish to make it generic like this:

def general_confusion_matrix(biopsy, column_name):
val=0
if biopsy==0:
    if column_name == 0:
        val = 0
    elif column_name == 1:
        val = 1
elif biopsy==1:
    if column_name == 1:
        val = 2 
    elif column_name == 0:
        val = 3
elif column_name == 2:
    val = 4
return val

so that I can apply it in this function something like this (this does not work).

def create_logit_value(df, name_of_column):
   df[name_of_column + '_concordance'] = df.apply(lambda : general_confusion_matrix('biopsy', name_of_column + '_category'), axis=1)

The issue seems to be that when you pass the columns in as df['biopsy'] you are passing a series to the general_confusion_matrix function rather than a value at each row and the conditional statements throw and the usual

   ('The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().', 'occurred at index 0')"

I have tried both map and apply but I am not sure how I can pass 2 arguments that refer to columns in my dataframe to the function in the lambda statement. I guess I could use map, but again, how do I pass the arguments through it. I apologise for writing 2 closely related questions but they are different.

jezrael · Accepted Answer · 2017-08-11 09:31:18Z

6

I think you are close:

df = pd.DataFrame({'biopsy_bin':[0,1,0,1,0,1],
                   'pioped_logit_category':[0,0,0,1,1,1],
                   'a_category':[0,0,0,1,1,1]})
print (df)


def create_logit_value(df, name_of_column):
    df[name_of_column + '_concordance'] = df.apply(lambda x: generate_confusion_matrix(x['biopsy_bin'], x[name_of_column + '_category']), axis=1)
    return (df)

create_logit_value(df, 'a')
create_logit_value(df, 'pioped_logit')

   a_category  biopsy_bin  pioped_logit_category  a_concordance  \
0           0           0                      0              0   
1           0           1                      0              3   
2           0           0                      0              0   
3           1           1                      1              2   
4           1           0                      1              1   
5           1           1                      1              2   

   pioped_logit_concordance  
0                         0  
1                         3  
2                         0  
3                         2  
4                         1  
5                         2

answered Aug 11, 2017 at 9:31

jezrael

868k103 gold badges1.4k silver badges1.3k bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

jezrael Over a year ago

Glad can help! Nice weekend!

Collectives™ on Stack Overflow

Pandas: passing multiple column names as arguments to a function with apply

1 Answer 1

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related