How to change abbreviated values in a DF using Pandas in Python

Question

I have the following problem, I have a df with more than 50 columns, in these columns, some have value of type "35,57B", "6,85T". How do I make iteratively and replace the examples cited above so that they stay as follows "3557000000","6850000000000"?

I tried something like

for col in df:
    df.col = (df.col.replace(r'[KMBT]+$', '', regex=True).astype(float) * df.col.str.extract(r'[\d\.]+([KMBT]+)', expand=False).fillna(1).replace(['K','M','B','T'], [10**3, 10**6, 10**9, 10**12]).astype(int))

But I was unsuccessful, and had the following mistake:

AttributeError: 'DataFrame' object has no attribute 'col'

Thanks for the answers with them it was possible to elaborate the solution. Which was the following:

for col in df.columns[1:]:    
    df[col] = (df[col].str.replace(',','.').replace('N/A','0').replace(r'[kMBT%]+$', '', regex=True).astype(float) * df[col].str.extract(r'[\d\.]+([kMBT%]+)', expand=False).fillna(1).replace(['k','M','B','T',"%"], [10**3, 10**6, 10**9, 10**12,1]).astype(float))
    df[col] = df[col].astype(str)
    df[col] = df[col].str.replace('.',',')

Answer 1

One option is to define a function, use list comprehension and iterate over df.columns:

def transform_str(value:str):
    return value.replace(",","").replace("B","000000").replace("T","0000000000")

for col in df.columns:
    df[col]=[transform_str(k) for k in df[col]]

Answer 2

In addition to the response of Lucas (that I even prefer), you can keep the basis of your code

Creating the test data frame

import pandas as pd
dados = ["35,57B", "6,85T"]
df = pd.DataFrame({'Dados': dados})
df

Displaying the data frame

     Dados
0   35,57B
1    6,85T

I added an extra replace and in regex I put a comma in place of the point, astype changed to int in both conditions

df['Dados'] = df['Dados'].str.replace(',','').replace(r'[KMBT]+$', '', regex=True).astype('int64') * df['Dados'].str.extract(r'[\d\,]+([KMBT]+)', expand = False).replace(['K','M','B','T'], [10**3, 10**6, 10**9, 10**12]).astype('int64')

If you want to save as string

df['Dados'] = df['Dados'].astype(str)

Exit:

0      3557000000000
1    685000000000000