String.replaceAll() is not working for some strings

Question

I am editing some email that got from tesseract ocr.

Here is my code:

 if (email != null) {
        email = email.replaceAll(" ", "");
        email = email.replaceAll("caneer", "career");
        email = email.replaceAll("canaer", "career");
        email = email.replaceAll("canear", "career");
        email = email.replaceAll("caraer", "career");
        email = email.replaceAll("carear", "career");
        email = email.replace("|", "l");
        email = email.replaceAll("}", "j");
        email = email.replaceAll("j3b", "job");
        email = email.replaceAll("gmaii.com", "gmail.com");
        email = email.replaceAll("hotmaii.com", "hotmail.com");
        email = email.replaceAll(".c0m", ".com");
        email = email.replaceAll(".coin", ".com");
        email = email.replaceAll("consuit", "consult");
    }
    return email;

But the output is not correct.

Input :

amrut=ac.hrworks@g mai|.com

Output :

lalcl.lhlrlwlolrlklsl@lglmlalil|l.lclolml

But when I assigned the result to a new String after every replacement, it works fine. Why continuous assignment in the same String is not working?

When I copy/paste that in (and fix the typo on line 5), it results in what seems to be a correct result - "amrut=ac.hrworks@gmail.com". — Ren, Feb 12 '13 at 05:45
Consider using String.replace instead of replaceAll. It does exactly what I believe you expected replaceAll to do. — Buhb, Feb 12 '13 at 06:58

Brian Roach · Accepted Answer · 2013-02-12T05:50:28.003

45

You'll note in the Javadoc for String.replaceAll() that the first argument is a regular expression.

A period (.) has a special meaning there as does a pipe (|) as does a curly brace (}). You need to escape them all, such as:

email = email.replaceAll("gmaii\\.com", "gmail.com");

edited Feb 12 '13 at 05:50

answered Feb 12 '13 at 05:44

Brian Roach

76,169
12
136
161

Do I have to escape | and { – Neeraj Feb 12 '13 at 05:50
4

Yes. Hence "You need to escape them all". – Brian Roach Feb 12 '13 at 05:52
1

I escaped | by \\| But still it is not getting replaced. Everything else works fine – Neeraj Feb 12 '13 at 06:05
4

You're using `String.replace()` there (I didn't even notice) which is *different*. It takes a `CharSequence` as the first argument (a `String`) and you don't escape there. – Brian Roach Feb 12 '13 at 06:18
though i was aware of it, it got skipped off my head ...Thank you for saving my day – Dila Gurung Feb 26 '18 at 14:09
the best way to escape all that needs to be escaped is `Pattern.quote(String)` – FiruzzZ Jun 27 '18 at 13:12

score 14 · Answer 2 · answered Feb 12 '13 at 05:52

14

(Is this Java?)

Note that in Java, replaceAll accepts a regular expression and the dot matches any character. You need to escape the dot or use

somestring.replaceAll(Pattern.quote("gmail.com"), "replacement");

Also note the typo here:

email = emai.replaceAll("canear", "career");

should be

email = email.replaceAll("canear", "career");

answered Feb 12 '13 at 05:52

David M. R.

313
1
7

Yes, Java. Thanks for mentioning that typo error also. I corrected – Neeraj Feb 12 '13 at 05:58

Sumit Singh · Answer 3 · 2013-02-12T10:28:16.110

You have to escape . by \\.like following :

if (email != null) {
    email = email.replaceAll(" ", "");
    email = email.replaceAll("caneer", "career");
    email = email.replaceAll("canaer", "career");
    email = email.replaceAll("canear", "career");
    email = email.replaceAll("caraer", "career");
    email = email.replaceAll("carear", "career");
    email = email.replace("|", "l");
    email = email.replaceAll("}", "j");
    email = email.replaceAll("j3b", "job");
    email = email.replaceAll("gmaii\\.com", "gmail.com");
    email = email.replaceAll("hotmaii\\.com", "hotmail.com");
    email = email.replaceAll("\\.c0m", "com");
    email = email.replaceAll("\\.coin", "com");
    email = email.replaceAll("consuit", "consult");
}
return email;

iTech · Answer 4 · 2013-02-12T06:33:29.723

6

By realizing that replaceAll() first argument is regex you can make your comparisons much less

For example you can check for possible misspellings of the word career by the following regex

email = email.replaceAll("ca[n|r][e|a][e|a]r", "career"));

edited Feb 12 '13 at 06:33

answered Feb 12 '13 at 06:11

iTech

18,192
4
57
80

1

Should that really be `[n|r]` and friends rather than `[nr]` & Co? – user Feb 12 '13 at 13:26
Yes you can, I thought the first one is explicit and more readable – iTech Feb 12 '13 at 16:23

score 5 · Answer 5 · answered Feb 12 '13 at 05:47

5

You are using some regex characters.

Please escape them using \ or by using Pattern.quote method

answered Feb 12 '13 at 05:47

Thihara

7,031
2
29
56

score 5 · Answer 6 · answered Feb 12 '13 at 05:47

I think you are not aware that first parameter of replaceAll is regex.

. , |, } might be interpreted in a different way from your expectation.

.   Any character (may or may not match line terminators)

http://docs.oracle.com/javase/6/docs/api/java/util/regex/Pattern.html

For space you better use

\s  A whitespace character: [ \t\n\x0B\f\r]

and escape other special characters with a leading \\

String.replaceAll() is not working for some strings

6 Answers6

Linked

Related