regexp：仅匹配指定符号内的单独数字

Question

I am using regular expression in java and can not figure out how to match only numbers inside predefined symbols.

here is example of one of my string:

some text [1] some and numbers 12, 14 and more text [11,1] another text [3,6, 7] and some more text [5;16]

Is it really possible to retrieve only numbers in square brackets? in this case: 1 11 1 3 6 7 5 16 This should not match any other symbols in square brackets.

I have already tried several options, including:

Pattern k = Pattern.compile("\\[(\\d+)\\]");

But this only gives works for one number and not others. I have tried to group them, like (?:\\\\[) some code inside (?:\\\\]) without any success.

Update

Workaround with grouping: Pattern k = Pattern.compile("\\\\[(\\\\d+)(?:.)?(\\\\d+)?\\\\]"); But produces the brackets and commas as output.

Answer 1

You can select those numbers, you can use the following regex:

(?:\G|\[)[,;\s]*(\d+)

It only selects the numbers.

See the demo

Explanation

• (?:\\G|\\[) matches the end position of the previous match or an opening bracket
• [,;\\s]* matches a comma, a semicolon or a space zero or multiple times
• (\\d+) captures the numbers

You'll have the numbers in the group 1.

Answer 2

Just to provide a different workaround that might come in handy when it gets even more complicated:

A much simplier way to achieve your goal is to separate the matching into two stages. First, match the bracket's occurences with

Pattern areas = Pattern.compile("\\[.*?\\]");

which will result in [1] [11,1] [3,6, 7] [5;16] .

Then for any of these use something like

Pattern numbers = Pattern.compile("\d+");

to match the actual numbers.

Answer 3

May be this one will help you.

Try it:

 Pattern k = Pattern.compile("\\[(\\d+)(;?)(\\d+)\\]");