Removing only certain types of characters regex

Question

Edit: Sorry, should have given more information

Is it possible to replace only the english text here?

{ "question": "本能", "answers":[ "ほんのうInstinct" ] },

I want it to just be { "question": "本能", "answers":[ "めがね" ] }, and keep the other english text.

Tried [A-Za-z]+ and it works for all english text, but I don't know how to select a certain section. I'm using find and replace on notepad+

@Laurianti That should be `.replace(/[A-Za-z]/g, '')`. [`[A-z]` matches more than just letters](http://stackoverflow.com/a/29771926/3832970). — Wiktor Stribiżew, Aug 10 '17 at 20:24
@WiktorStribiżew I'm using Notepad+ find and replace at the moment. Each line is different — PierceJ, Aug 10 '17 at 20:26
@PierceJ Then, really, you may use that simple regex, `[A-Za-z]+`. Try. — Wiktor Stribiżew, Aug 10 '17 at 20:27
@WiktorStribiżew that works perfect, but how would I include the quotation and the brace so that it only works with this part of the line? — PierceJ, Aug 10 '17 at 20:30
Wait, what do you mean? I thought you need to remove the letters anywhere in the document. — Wiktor Stribiżew, Aug 10 '17 at 20:31
The full line is this: { "question": "本能", "answers":[ "ほんのうInstinct" ] }, sorry I should have mentioned it. I only want to remove the english after answers — PierceJ, Aug 10 '17 at 20:33
Please edit the question mentioning what you really need and what you tried and what is wrong with the current approach. Also, add a relevant tag. — Wiktor Stribiżew, Aug 10 '17 at 20:34
The real problem is the fact it is a JSON file. You should really think of using a JSON parser here. If you think you may use a regex, then please define the *end* boundary for the pattern. The starting boundary is `"answers":`, but when should it stop matching? — Wiktor Stribiżew, Aug 10 '17 at 20:37
You have a space between `[` and `"`, so use `(?:\G(?!^)|"answers"\s*:\s*\[\s*")[^"]*?\K[a-zA-Z]+`. I do not know what the indentation is there, in your file, just add `\s*` anywhere where spaces may occur. Besides, it is not clear if you have just "trailing" English letters, or there can be `"SOMEめがねwordsがねhere"` — Wiktor Stribiżew, Aug 10 '17 at 20:51

score 1 · Answer 1 · answered Aug 10 '17 at 20:28

1

this is your regex: /[A-Za-z]*/g

var s = '":[ "めがねGlasses" ] },'

console.log(s.replace(/[A-Za-z]*/g, ''))

answered Aug 10 '17 at 20:28

jdubjdub

613
7
21

Ulysse BN · Answer 2 · 2017-08-10T20:34:02.380

1

You can replace everything that matches /[a-z]/gi by an empty string.

The flags g and i respectively means that you want to match multiple times and you want to be case insensitive.

You also could use \w shortcut which stand for word character, but this will also match [0-9_] which you may not want to delete.

Here is a working exemple of the regex.

edited Aug 10 '17 at 20:34

answered Aug 10 '17 at 20:28

Ulysse BN

10,116
7
54
82

score 1 · Accepted Answer · answered Aug 10 '17 at 20:58

You may use

(?:\G(?!^)|"answers"\s*:\s*\[\s*")[^"]*?\K[a-zA-Z]+

and replace with an empty string.

Details:

(?:\G(?!^)|"answers"\s*:\s*\[\s*") - the end of the previous successful match (\G(?!^)) or (|) "answers" substring, 0+ whitespaces, :, 0+ whitespaces, [, 0+ whitespaces, " (see "answers"\s*:\s*\[\s*")
[^"]*? - any 0+ chars other than ", as few as possible (as *? is a lazy quantifier)
\K - a match reset operator discarding all text matched so far from the match buffer
[a-zA-Z]+ - 1 or more ASCII letters

Test:

Removing only certain types of characters regex

3 Answers3