Extract substring from these strings using regex (Java)

Question

I am trying to extract the TPS number from the following strings using Java - the strings will be read from a file and so the strings can appear in any order (not known in advance), e.g. I wont know which of the strings I am dealing with - it could be either of these two:

Testing performance TPS..  ok. (795 TPS recorded for run)

Testing performance TPS..  warning: TPS seems low - it was 10 TPS and I expected to achieve over 50

E.g. for the first string I would want the number 795, and for the second string I would want the number 10.

Does anyone know how to do this with regex or similar using Java?

Many thanks

The regex to get the first number in a text line is something like this: ^.*([0-9]*). — Viruzzo
– Viruzzo, Commented Dec 12, 2011 at 14:45
This is quite a flaky way of getting these numbers - are you sure you can't directly hook into the source of the file, and whatever produces these lines in the file? It would be a far better way to get the data you want. If not, the regexp in answers below will do the trick, but make sure you validate (at runtime) that you're reading a line that has the correct (i.e. expected) format — Guillaume
– Guillaume, Commented Dec 12, 2011 at 14:50

ewan.chalmers · Accepted Answer · 2011-12-13 09:59:42Z

4

You need to find the first group of number characters in the input. The number is terminated by a space.

You can use this regex:

    String regex = "[^\\d]+(\\d+) .*";

The number is captured in group one ($1).

Here is a simple test:

public static void main(String[] args) throws Exception {

    String[] lines = {
        "Testing performance TPS..  ok. (795 TPS recorded for run)",
        "Testing performance TPS..  warning: TPS seems low - it was 10 TPS and I expected to achieve over 50"
    };

    String regex = "[^\\d]+(\\d+) .*";
    Pattern p = Pattern.compile(regex);
    for (String s: lines) {
        Matcher m = p.matcher(s);
        if (m.matches()) {
            System.err.println(m.group(1));
        }
    }
}

The output is:

795
10

edited Dec 13, 2011 at 9:59

answered Dec 12, 2011 at 14:44

ewan.chalmers

16.3k47 silver badges62 bronze badges

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

2 Comments

Alan Moore Over a year ago

Your use of groupCount() is incorrect. It just tells you how many capturing groups there are in the regex. It doesn't say anything about what was actually matched. To find out if group #1 participated in the match, you use if (m.group(1) != null) or if (m.start(1) != -1).

ewan.chalmers Over a year ago

@Alan Thanks for pointing that out, I didn't know. So really you only need to check that Matcher.matches() before attempting to access groups. If the pattern matches then you can lookup all your groups, though some might be null or empty.

OpenSauce · Accepted Answer · 2011-12-12 14:46:38Z

3

If you're always looking for an integer followed by the string "TPS" you can do

"(\\d+) TPS"

But you'd better be sure it will always be in this format - it would be better to modify the output format, if that's possible.

answered Dec 12, 2011 at 14:46

OpenSauce

8,6731 gold badge26 silver badges30 bronze badges

3 Comments

Peter Perháč Over a year ago

excellent. in java, just use if(matcher.groupCount()>=1) { String groupStr = matcher.group(1); ... } This will give you the first captured group. Zero-th group would be the whole match, i.e. "795 TPS", first group will be just "795"

Alan Moore Over a year ago

@PeterPerháč: That's not what groupCount() is for. See my comment to sudocode's answer for details.

Peter Perháč Over a year ago

you win :-) great and simple answer... I was of course thinking about positive lookbehinds and stuff.. but those are zero-width and don't capture... and then you posted this and I was like d'oh!

Stephen C · Accepted Answer · 2011-12-12 15:07:18Z

1

This regex should do the trick:

    ^[^0-9]*([0-9]+).*$

It matches any line that contains a number, and extracts the first number in the line.

However, it is not really possible to generalize from just these two examples. For instance you don't show us examples that the regex shouldn't match.

I agree with the comment that say that this is a flakey way to extract information. Unless you are very sure of your input text, there is always a possibility that you will encounter a different form that the regex doesn't cope with; e.g. that matches when it shouldn't or vice versa.

edited Dec 12, 2011 at 15:07

answered Dec 12, 2011 at 14:47

Stephen C

723k95 gold badges849 silver badges1.3k bronze badges

Collectives™ on Stack Overflow

Extract substring from these strings using regex (Java)

3 Answers 3

2 Comments

3 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

3 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related