Replace strings from column with unique random strings

user_7338182 0
Viewed 465

I have a a csv that has multiple columns, one of these columns consists of strings.

I start with just reading the csv file and then just using two columns 

df = pd.read_csv("MyDATA_otherstring.csv", usecols=["describe_file", "data_numbers"])

This is the output 

    describe_file   data_numbers
0   This is the start of the story  7309.0
1   This is the start of the story  35.0
2   This is the start of the story  302.0
3   Difficult part  7508.5
4   Difficult part  363.0

In around 10k rows, there are around 150 unique strings. These strings appear multiple times within the file.

My goal Filter by the first string example 'This is is the start of the story' and replace it with a random string.

I want to run over all the strings in that column and replace them with unique strings

I have looked into the random library and some questions that have been asked here, unfortunately I have not found anything that would help me.

pandas csv random python string
2 Answers

This is your example:

import pandas as pd
import numpy as np
from string import ascii_lowercase

df = pd.DataFrame([['This is the start of the story']*3 + ['Difficult part']*2, 
    np.random.rand(5)], index=['describe_file', 'data_numbers']).T

                    describe_file data_numbers
0  This is the start of the story     0.825913
1  This is the start of the story     0.704422
2  This is the start of the story      0.91563
3                  Difficult part     0.192693
4                  Difficult part     0.795088

This is how you can do it:

df.describe_file = df.join(df.groupby('describe_file')['describe_file'].apply(lambda x:
    ''.join(np.random.choice(list(ascii_lowercase), 10))), \
    on='describe_file', rsuffix='_NEW')['describe_file_NEW']

The result:

  describe_file data_numbers
0    skgfdrsktw     0.204907
1    skgfdrsktw     0.399947
2    skgfdrsktw     0.990196
3    rziuoslpqn     0.930852
4    rziuoslpqn     0.210122
user_109083750

The previous answer by @Nicolas Gervais is fine, but after reading several times the question I interpret that the question is to replace 'This is the part of the story' by a random string, but leave the rest 'Difficult part' as it is. The following command including .replace() statement is doing that.

df['describe_file'].apply(lambda x: x.replace('This is the start of the story', ''.join(np.random.choice(list(ascii_lowercase), 10)))) 

0        glhrtqwlnl
1        qxrklnxhoj
2        kszgtysptj
3    Difficult part
4    Difficult part
Name: describe_file, dtype: object
user_40086570
Related