Regex to parse file name in Java

Question

I am trying to parse a file name according to a given pattern but not able to perfect the match. Here is a sample file name:

CRS-ISAU-RPV#3430_Dedalus_Conc.ok.erto_AOTreviglio.doc

And here are my requirements:

til the character # the file name can contain anything, after #, i have to find character _ or the character - to separate a string. The string in between the character(optionally _ or - - but not both) can contain any other character. So eventually after the character # i must have exactly three (3) _ or - characters combined. The string should end with .doc or .docx or .odt but NOT .ok.doc or .ok.docx or .ok.odt.

Here is what i tried:

(.*)#([^_-]+)[_-]([^_-]+)[_-]([^_-]+)[_-]([^_-]+)\.[doc|odt|docx].*(?<!\.ok)$

But this forces me to end the string with .doc.ok or .docs.ok or .docx.ok and actually i want to retain the file extension at the end.

If i try this:

(.*)#([^_-]+)[_-]([^_-]+)[_-]([^_-]+)[_-]([^_-]+)\..*(?<!ok\.[doc|odt|docx])$

it wont work.

Any help would be appreciated. Thank you :)

The syntax [doc|odt|docx] doesn't do what you appear to think it does. Try replacing the [] with a non-capturing group: (?:) — jsheeran
– jsheeran, Commented Apr 22, 2016 at 10:39
Try "^([^#]*#[^-_]*)[-_](.*)$(?:(?<=(?<!\\.ok)\\.docx$)|(?<=(?<!\\.ok)\\.doc$)|(?<=(?<!\\.ok)\\.odt$))" — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Apr 22, 2016 at 10:51
@WiktorStribiżew ocpsoft.org/tutorials/regular-expressions/… this link throws an error for what you are suggesting. — fresh learner
– fresh learner, Commented Apr 22, 2016 at 10:57
No, there, you need to insert a single-backslashed version: (?m)^([^#]*#[^-_]*)[-_](.*)$(?:(?<=(?<!\.ok)\.docx$)|(?<=(?<!\.ok)\.doc$)|(?<=(?<!\.ok)\.odt$)) - or even simpler (?m)^([^#]*#[^-_]*)[-_](.*)$(?<=(?<!\.ok)\.(?:docx?|odt)$) works there. — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Apr 22, 2016 at 11:02

Wiktor Stribiżew · Accepted Answer · 2016-04-22 11:16:23Z

2

It seems you can use

^([^#]*#[^-_]*)[-_](.*)$(?<=(?<!\.ok)\.(?:docx?|odt)$)

Explanation:

^ - start of string (not necessary when used with .matches(), but not harmful)
([^#]*#[^-_]*) - Group 1: any 0+ characters other than # ([^#]*) followed with # and then any 0+ characters iother than - and _ (with [-_])
(.*)$ - match 0+ characters other than a newline (since DOTALL mode is not specified) up to the end of string BUT...
(?<=(?<!\.ok)\.(?:docx?|odt)$) - after reaching the end, check if there is .doc or .docx or .odt at the end (see (?<=\.(?:docx?|odt)$)) that are not preceded with .ok (see (?<!\.ok)). In PCRE, these conditions should be split, Java regex seems to cope with alternations inside the lookbehind.

A lookahead-based alternative:

^([^#]*#[^-_]*)[-_](?=.*(?<!\.ok)\.(?:docx?|odt)$)(.*)$

See the regex101 demo. It is the same, but all the end-of-string checks are done after matching the - or _.

See the Java demo:

List<String> strs = Arrays.asList("CRS-ISAU-RPV#3430_Dedalus_Conc.ok.erto_AOTreviglio.doc",
            "CRS-ISAU-RPV#3430_Dedalus_Conc.ok.erto_AOTreviglio.docx",
            "CRS-ISAU-RPV#3430_Dedalus_Conc.ok.erto_AOTreviglio.odt",
            "CRS-ISAU-RPV#3430_Dedalus_Conc.ok.erto_AOTreviglio.ok.docx",
            "CRS-ISAU-RPV#3430_Dedalus_Conc.ok.erto_AOTreviglio.ok.odt"
        );
for (String str : strs) {
    System.out.println("----------\nMatching: " + str);
    Matcher m = Pattern.compile("^([^#]*#[^-_]*)([-_])(.*)$(?<=(?<![.]ok)[.](?:docx?|odt)$)").matcher(str);
    if (m.matches()) {
        System.out.println(m.group(1));
        System.out.println(m.group(2));
        System.out.println(m.group(3));
    } else { System.out.println("No match"); }
}

edited Apr 22, 2016 at 11:16

answered Apr 22, 2016 at 11:10

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

fresh learner Over a year ago

Appreciate your answer. It misses one thing though in the question: " i must have exactly three (3) _ or - characters combined." Which means, collectively there has to be 3 _ or - Summing up the two characters, they must be as a whole 3. And in between these characters will be strings of any sort BUT not containing #. The sample file name(String) shows that there are 3 _ but you can replace any _ with - and yet it should be a match.

Wiktor Stribiżew Over a year ago

Maybe this? ^([^#\n]*#[^-_\n]*)[-_](?=(?:[^-_\n]*[_-]){3}[^-_\n]*$)(?=.*(?<!\.ok)\.(?:docx?|odt)$)(.*)$ (for testing) and "^([^#]*#[^-_]*)[-_](?=(?:[^-_]*[_-]){3}[^-_]*$)(?=.*(?<![.]ok)[.](?:docx?|odt)$)(.*)$" (for using in code).

fresh learner Over a year ago

I solved it with the help of your answer. I will mark your answer as correct :) Thanks a lot!

Collectives™ on Stack Overflow

Regex to parse file name in Java

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related