Pandas fillna the same value in a column, but different per column-CodePudding

I have a dataframe that contains NA's, I need to fill the NA's with some value, however I need the same value per column, but every column must be different. I cannot use a dictionary for column to NA value mapping as I don't know how many columns due to the data being imported from an API.

I am using python 3

sample of DF and expected output:

DF:

Expected output:

CodePudding user response：

You can craft a dictionary with replacement values for each column containing NAs (here I'm using ascii uppercase letters, but you can use whatever source):

s = df.isna().any()
d = dict(zip(s[s].index, ascii_uppercase))
# {'A': 'A', 'B': 'B', 'D': 'C', 'E': 'D'}

df = df.fillna(d)

output:

   a  b  c  d  e
0  -  -  -  -  D
1  A  B  -  -  -
2  A  -  -  C  D

used input:

     a    b  c    d    e
0    -    -  -    -  NaN
1  NaN  NaN  -    -    -
2  NaN    -  -  NaN  NaN

alternative method to craft the dictionary

s = df.isna().any()
d = {k: v*3 '|' for k,v in zip(s[s].index, ascii_uppercase)}
# {'a': 'AAA|', 'b': 'BBB|', 'd': 'CCC|', 'e': 'DDD|'}
df = df.fillna(d)

output:

      a     b  c     d     e
0     -     -  -     -  DDD|
1  AAA|  BBB|  -     -     -
2  AAA|     -  -  CCC|  DDD|

CodePudding user response：

IIUC:

maps = {col: i for i, col in enumerate(df.columns)}
df = df.fillna(maps)