Java Regex Metacharacters

Question

I found this thread and one of users on it posted the following line of code:

String[] digits2 = number.split("(?<=.)");

I have consulted a couple of sources- like 1 and 2-to decipher what this code mean but I can't figure it out. Can anybody explain what the argument in the split() method means?

Edit: To anyone who has the same question as I had, here's another helpful link

Sergey Kalinichenko · Accepted Answer · 2013-05-20 02:06:03Z

4

This is a positive lookbehind. The overall expression means "after any character, but without capturing anything". Essentially, if the string looks like

ABC

then the matches would occur at |, between the characters.

A|B|C|

answered May 20, 2013 at 2:06

Sergey Kalinichenko

729k85 gold badges1.2k silver badges1.6k bronze badges

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

3 Comments

Haque1 Over a year ago

I know that "." matches any character except for line breaks, but what does "<=" mean in this context?

Sergey Kalinichenko Over a year ago

@Haque1 That's a sequence of metacharacters that tells the expression that the "." should not capture anything. The engine needs to see that a character is there, but it should not remove that character from the stream.

Sergey Kalinichenko Over a year ago

@Haque1 This is an unusual use of lookbehind. A more common way is like this: (?<=tag:)"[^"]*". This expression matches a quoted string only when it is preceded by a tag: string.

Explosion Pills · Accepted Answer · 2013-05-20 02:07:10Z

1

.split("") (on an empty string/pattern) will match the empty string at the start of the regex. This is an additional empty string character that is undesirable. (?<=.) is a zero-width assertion (does not consume any characters) that matches the zero-width space followed by any character (followed by because it is a lookbehind). This splits on the empty string between each character, but not the empty space between the first character and the start of the string.

answered May 20, 2013 at 2:07

Explosion Pills

192k56 gold badges341 silver badges417 bronze badges

3 Comments

Cairnarvon Over a year ago

Zero-width space is the wrong term to use here; that's a different thing.

Explosion Pills Over a year ago

@Cairnarvon thanks for pointing this out ... which term would you prefer?

Cairnarvon Over a year ago

I want to say "character boundary", but that's not strictly right either. I'm not sure, to be honest.

Collectives™ on Stack Overflow

Java Regex Metacharacters

2 Answers 2

3 Comments

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related