Converting perl regex to Java

Question

I have a piece of regex I've been using in Perl to parse a large text file containing every stored procedure in an Oracle database. It looks like this:

/create\s+(proc(edure)?|function)\s+\[*(dbo)?\]*\.*\[*(\w+)/i

I'm not too familiar with Perl though, so would ideally like to convert this to Java. I've tried running it through some converters, which produces:

"/create\\s+(proc(edure)?|function)\\s+\\[*(dbo)?\\]*\\.*\\[*(\\w+)/i"

However, this doesn't seem to match any correct input. Can anyone give me some pointers in converting this correctly to Java?

Sample Data:

SET QUOTED_IDENTIFIER ON 
GO
CREATE PROCEDURE AddNewTc                       

@TCN NVARCHAR(100),                
@TCM NTEXT,                
@TCOLE IMAGE = NULL,                
@BYSPID INT = 0,

Only the "CREATE PROCEDURE" line should be matched.

can you please add expected input/output to help us translating this regex? i don't know Perl ReGex syntax so this can be helpful — Supamiu, Oct 29 '14 at 09:43
have you double-slash those slash? java string slash need escaping — J-16 SDiZ, Oct 29 '14 at 09:43
@J-16SDiZ: I did, but behind the scenes SO is playing with how both regex are represented. I'm not sure how to fix it. — Andrew Martin, Oct 29 '14 at 09:46
@AndrewMartin, use 4 spaces instead of > to get proper display of the double \'s — srbs, Oct 29 '14 at 10:16

score 1 · Answer 1 · answered Oct 29 '14 at 10:15

The converter works just fine, but you will need to change a couple of things.

/create\\s+(proc(edure)?|function)\\s+\\[*(dbo)?\\]*\\.*\\[*(\\w+)/i

Java does not have inline-regexes like you do in perl & javascript. You will need to use the Pattern & Matcher API for regex support.
Due to the lack of inline-regex, pattern modifiers are done separately with flags.

Applying step 1 (remove prepended / and appended /) and step 2 (replace /i with Pattern.CASE_INSENSITIVE) you would get something like:

 Pattern p = Pattern.compile("create\\s+(proc(edure)?|function)\\s+\\[*(dbo)?\\]*\\.*\\[*(\\w+)", Pattern.CASE_INSENSITIVE);
 boolean matches = p.matcher(str).matches();

`find()` is more appropriate here. – nhahtdh Oct 29 '14 at 10:47 — nhahtdh, Oct 29 '14 at 10:47

Supamiu · Answer 2 · 2014-10-29T10:17:36.457

0

Try with this one:

"create\\s+((procedure)?|(function))?\\s+\\([(dbo)?\\]\\.\\[)?(\\w+)"

It should work, but you need to add upper case if you need to match some, or add a "i" flag to your regex to ignore case.

in Java, ecape character \ must be escaped with another \ to escape in regex syntax (too much escape !)

DEMO

edited Oct 29 '14 at 10:17

answered Oct 29 '14 at 09:55

Supamiu

8,501
7
42
76

I think you're having the issue I was having. Adding the extra slashs on SO doesn't show. – Andrew Martin Oct 29 '14 at 10:00
Oh, in fact i typed them, but need to add a code area to show them, gonna fix it now – Supamiu Oct 29 '14 at 10:17

score 0 · Answer 3 · answered Oct 29 '14 at 12:53

0

The JRegex project (which I am neither affiliated with nor have been a user of) might interest you. It claims to provide a Java library to handle perl regular expressions.

answered Oct 29 '14 at 12:53

collapsar

17,010
4
35
61

Converting perl regex to Java

3 Answers3