How to replace multiple values in matching pattern from string

Question

I want to search a pattern in a string and then again search some invalid character in matching pattern and then remove them or replace with some valid characters.

I have some sample dictionaries eg. sample_dict = {"randomId":"123y" uhnb\n g", "desc": ["sample description"]}

In this case I want to find the value of a dictionary let say "123y" uhnb\n g" and then remove invalid characters in it such as (", \t, \n) etc.. what I have tried is stored all the dictionaries in a file then read file and matching pattern for dictionary value, but this gives me a list of matching pattern, I can also compile these matches but I am not sure how to perform replace in original dictionary value so my final output will be: {"randomId":"123y uhnb g", "desc": ["sample description"]}

pattern = re.findall("\":\"(.+?)\"", sample_dict)

expected result:

{"randomId":"123y uhnb g", "desc": ["sample description"]}

actual result:

['123y" uhnb\n g']

Possible duplicate of [Parse JSON in Python](https://stackoverflow.com/questions/7771011/parse-json-in-python) — miken32, Apr 20 '19 at 05:48
@miken32: I can use json parser but in that case as well I need to remove those invalid characters else it won't work, so in order to remove those characters I am using regex. — Pradeep, Apr 20 '19 at 06:06
How did you end up with this strange `sample_dict` to begin with? Perhaps you can avoid that already earlier so that you do not need to replace or remove the strange characters. — JohanL, Apr 20 '19 at 06:32
Why are you using re.findall instead of re.sub (in some capacity)? — FailSafe, Apr 20 '19 at 06:40
Secondly, why are you passing a dictionary to findall as opposed to a string? Are you implying that you convert the regex to a string, or are you inserting the randomid key's value? — FailSafe, Apr 20 '19 at 06:42
As suggested, using re.sub is easier to replace non-alphanumeric characters, check my answer below to see how! @Pradeep — Devesh Kumar Singh, May 03 '19 at 12:39

score 1 · Answer 1 · answered Apr 20 '19 at 06:59

You can just substitute non-alphanumeric characters in your value using re.sub as below

dct = {"randomId":"123y uhnb\n g", "desc": ["sample description"]}
import re

for key, value in dct.items():
    val = None
    #If the value is a string, directly substitute
    if isinstance(value, str):
       val = re.sub(r"[^a-zA-Z0-9 ]", '', str(value))
    #If value is a list, substitute for all string in the list
    elif isinstance(value, list):
       val = []
       for item in value:
           val.append(re.sub(r"[^a-zA-Z0-9]", ' ', str(item)))
    dct[key] = val

print(dct)
#{'randomId': '123y uhnb g', 'desc': ['sample description']}

How to replace multiple values in matching pattern from string

1 Answers1