Conditional lambda apply across dataframe based on list equality-CodePudding

I have a dataframe df whos columns contain lists of strings

df =      A               B 
        ['-1']    , ['0','1','2']
     ['2','4','3'],     ['2']
      ['3','8']   ,    ['-1']

I want to get the length of all the lists except the ones that are ['-1'] for the lists that are ['-1'] I want them to be -1

Expected output:

df = A   B 
    -1,  3
     3,  1
     2, -1

I've tried

df.apply(lambda x: x.str.len() if not x == ['-1'] else -1)

and got the error ('Lengths must match to compare', (132,), (1,))

I have also tried

data_copy[colBeliefs] = data_copy[colBeliefs].apply(lambda x: x.str.len() if '-1' not in x else -1)

but this produces the wrong output where ['-1'] becomes 1 rather than -1

I'm not sure how I can apply functions to a dataframe based on the whether an entry in a dataframe is equal to a list, or whether an item is in a list.

EDIT: Output of df.head().to_dict()

{'A': {0: ['-1'],
       1: ['2','4','3'],
       2: ['3','8']},
 'B': {0: ['0','1','2'],
       1: ['2'],
       2: ['-1']}}

CodePudding user response：

You could do:

df.applymap(lambda x: -1 if (ln:=len(x)) == 1 and x[0] == '-1' else ln)

   A  B
0 -1  3
1  3  1
2  2 -1

Edit:

If yousing python < 3.8 Use the following:

df.applymap(lambda x: -1 if len(x) == 1 and x[0] == '-1' else len(x))

CodePudding user response：

The comparison doesn't work for lists, but it works for tuples. So you could convert them to tuples to do avoid the ValueError:

s = df.stack()
out = s.str.len().mask(pd.Series(map(tuple, s), index=s.index)==('-1',), -1).unstack()

Output: