Java RegEx Syntax with Word Boundaries?

Question

Firstly, my syntax will not be part of a script as such but it will be parsed via a form input--so any 'existing' solution pointing to Java code will not apply per se.

Okay, so here is what I need to do: I need to be able to input a term like:

'This is your airport and this is your car.' into an input field in such a way that only the word 'airport' or 'airports' to be matched. So nothing like '99airport' or 'airport99' should be matched. And I am close!

(?i).*\bair[port|ports].*

If I input the above as RegEx in a test site:

http://www.ocpsoft.org/tutorials/regular-expressions/java-visual-regex-tester/#!;t=123-45-6789%0A9876-5-4321%0A987-65-4321%20(attack)%0A987-65-4321%20%0A192-83-7465&r=(%3Fm)%5E(%5Cd%7B3%7D-%3F%5Cd%7B2%7D-%3F%5Cd%7B4%7D)%24&x=Found%20good%20SSN%3A%20%241

then, indeed, '99airport' does not match because of the beginning use of the Word Boundary identifier \b ; However, I don't know how to put the \b around the ending of the word so that 'airport99' also does not match. I have tried a few things but no luck. I think it is the syntax to be put around the [] which needs to be figured out.

And please don't pay too much attention to what needs to be matched or not--these are just random words. Currently, if my input has 'airport99' it does get matched but it shouldn't if I can figure out a solution.

Thanks!

Just use (?i).*\bairports?\b.*. In Java, String patrn = "(?i).*\\bairports?\\b.*"; — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Nov 16, 2015 at 14:31
Thanks. That would work except I also need the possible combinations inside the [ ]; you see, in real life, I am trying to catch words like 'tornado' or 'Tornado' or 'ToRNADOES' while preventing to catch 'FCGTornadoTeam' . — IrfanClemson
– IrfanClemson, Commented Nov 16, 2015 at 14:33
You are mistaking groups for character classes. Let me explain in an answer. — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Nov 16, 2015 at 14:36

Wiktor Stribiżew · Accepted Answer · 2015-11-16 14:38:41Z

2

I see you are using mather.matches to check for a word inside the input string. That is why you need the .* before and after a keyword. Since the text is coming from an input field, you do not need to match newline symbols, and no need in (?s) singleline/dotall modifier.

However, you mistake character classes ([...]) with groups ((...)). Character classes match 1 character. For example, [port|ports] matches 1 character, either p, o, r, t, |, or s. Groups can be used to match specific sequences of symbols. E.g. (port|ports) will match either port or ports.

Thus, in your case, you can use

(?i).*\bairports?\b.*

or - less effecient -

(?i).*\bair(port|ports)\b.*

In Java, String patrn = "(?i).*\\bairports?\\b.*";

answered Nov 16, 2015 at 14:38

Wiktor Stribiżew

631k41 gold badges502 silver badges632 bronze badges

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

2 Comments

IrfanClemson Over a year ago

Yes, (?i).*\bair(port|ports)\b.* does it! My problem was using [ ] instead of () to catch port|ports. I will see if this will work inside the actual program or not--it should. Thanks for your help and the explanation. New to RegEx and not really a Java person.

Wiktor Stribiżew Over a year ago

I really would like to stress the fact that just using ports? (where s? matches 1 or 0 s) is a better alternative to (port|ports). For country/countries, you can't escape using an alternative groups, but wherever possible, just use optional quantifier ? (1 or 0 occurrences). It will involve much less backtracking, and improve performance. Well, if your input strings are short, that should not be a problem. Just best practices are best practices :)

Little Santi · Accepted Answer · 2015-11-16 14:38:39Z

1

This expression should match your requirements:

(?i)\\b(air)?port\\b

It does match "port" and "airport" but does not match "99port" nor "port99" nor "99airport" neither "airport99".

If a more generic expression is needed, this one should match any word starting with "air" plus some other (optional) letters, but no digits or punctuation symbols:

(?i)\\b(air)?[a-z]*\\b

edited Nov 16, 2015 at 14:38

answered Nov 16, 2015 at 14:14

Little Santi

8,8532 gold badges22 silver badges50 bronze badges

4 Comments

IrfanClemson Over a year ago

I admit maybe I didn't phrase the question 100% correctly--if I could just prevent both '99airport' and 'airport99' to match while still retaining the possible words inside the [ ] then that would be all I need. I still need to account for what is inside [ ] and not sure your solution accounts for that. Apologies.

Little Santi Over a year ago

Hehe. Yes, it happens :-) I suggest you include in your question several examples of what should be matched and other examples of what should not be matched.

IrfanClemson Over a year ago

Thanks for understanding. Rephrased the question.

IrfanClemson Over a year ago

Thanks! I think stribizhev's Answer perfectly matches the requirements; I was close but was using [ ] instead of () to catch the port or ports possibilities.

Collectives™ on Stack Overflow

Java RegEx Syntax with Word Boundaries?

2 Answers 2

2 Comments

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related