CODEWITHSUNDEEP
pythonpandas
kitchen800
kitchen800

Reputation: 227

Changing Dataframe type with an exception

I have a few pandas dataframe that consist of 1 column and 5 rows, each dataframe contains 3 values, "1", "0", and "P". These data frames are of type "object".

I am running into problems when i am adding data frames together. il give you an example:

df1 = [ 1, P, 0, 1, 1]
df2 = [1, P, 1, 1,1]
df1 + df2 = [11, PP, 01, 11,11]

Note: A "P" will alway be in the same location in all data frames. so when adding together data frames i could get "PPPPP" for example.

In the example above i am trying to get the output to be:

df1 + df2 =  [2, P, 1, 2, 2]

What i am thinking is to use pandas to make each dataframe column numeric and then i should be able to add them together.

pd.to_numeric(df1["my df title"], errors='coerce')

the problem with this is that the "P" values cannot be converted to numeric. how to i make everything numeric except the p values?

what my real dataframe looks like when I add df1 + df2 + df3 + df4:

0     1111
1     PPPP
2     1111
3     1111
4     PPPP

Upvotes: 0

Views: 68

Answers (1)

Ynjxsjmh
Ynjxsjmh

Reputation: 30032

Just convert each value to expected output with if-else.

import pandas as pd

df1 = pd.DataFrame({'col': [ '1', 'P', '0', '1', '1']})
df2 = pd.DataFrame({'col': [ '1', 'P', '1', '1', '1']})

s = df1['col'] + df2['col']

s = s.map(lambda x: 'P' if 'P' in x else sum(map(int, list(x))))

print(s)

0    2
1    P
2    1
3    2
4    2
Name: col, dtype: object

Upvotes: 2

Related Questions