正则表达式性能问题

Question

I have to construct a regex for angularjs ng-pattern attribute. The regex has to validate a text, not each line or some pieces. This is the format i have to validate: some text with spaces and punctuation:digits.2digits

For example this text is valid:

name:1234.32 name:234.43<br>
name:43.22

For example this text is not valid because one group starts with ":":

name:1234.32 :234.43<br>
name:43.22

For example this text is not valid because last group doesn't end with 2 decimals:

name:1234.32 name:234.43 name:43.2

I have build some regex but in case i have a bigger text, it fails with timeout . It fails only when last group is invalid. Here is an example.
This is my regex:

^(([\S\s]+)\s*:\s*([0-9]+[.][0-9]{2})\s*)+$

Can you help me optimize this regex?

Answer 1

Your own regex has multiple following quantifiers with no limitation on matching characters that causes engine to encounter a catastrophic backtracking almost on failure on large data.

You'd go with a regex like this:

^(([^:]*)\S\s*:\s*\d+\.\d{2}(?!\d))*$

Live demo

Breakdown:

^   # Assert beginning of input string
(   # Construct a group #1
    ([^:]*) # Construct group #2, match anything except `:`
    \S\s*:\s*\d+\.\d{2}(?!\d)   # Match a valid sequence
)*  # Repeat #1 as much as possible
$   # End of input string

Answer 2

Try this ^(([\\S\\s]+)\\s*?:\\s*?([0-9]+[.][0-9]{2})\\s*?)+$

But I recommend if you can, split text on lines and evaluate each one, and stop at line that not complain the test.