Replace int values in string with letters in a column using Python

user_11050535 0
Viewed 236

I currently have a csv file. The data originally is derived from PDF and doing a further analysis on the data, There are certain rows where the extracted data contains letters in place of numbers,

I need instead of numbers the letters of the variables. So trying to replace the int values by the letters

Such as in the following example:

col_state

2567i
28981
2534s
0123o

in the above table i am looking out to replace (i=1, s=5, o=0)

Expected Output:

col_state

25671
28981
25345
01230

What i have tried so far:

import re
chars_to_remove = ['i', '1', 's', '5', '']
regular_expression = '[' + re.escape (''. join (chars_to_remove)) + ']'

df['col_state'].str.replace(regular_expression, '', regex=True)

print(df['HSN_Code'])

So I have no clue how to handle this problem :(

pandas python
2 Answers

You might use translate method as follows

import pandas as pd
data = pd.Series(["2567i","28981","2534s","0123o"])
t = str.maketrans("iso","150")
data = data.str.translate(t)
print(list(data))

output

['25671', '28981', '25345', '01230']

Explanation: translate is useful when you need replace single characters using other single characters, str.maketrans when using in 2-argument form create replacement table so i-th element of 1st argument is replaced using i-th element of 2nd argument (arguments must be equal length) which is then usable in translate. translate is method of strs and can be used without pandas.

user_107859750

Try with:

repl = {'i':'1', 's':'5', 'o':'0'}
df['col_state'] = df['col_state'].replace(repl, regex=True)

Output:

  col_state
0     25671
1     28981
2     25345
3     01230
user_42384080
Related