Question

I’m not so good with regex, and need some help.

I have a string similar to the following:

[{type='(type here)', field='(field here)', value='(value here)'},{...},...,{...}]

I am trying to match it with the following regex:

^\[(\{type=\'(.*)\', field=\'(.*)\', value=\'(.*)\'\},*)*\]$

But it isn’t matching. I then debugged. Here is the regex I used for debugging:

\[(\{(.*)\}\]

Here is the string:

[{type='cost', field='flag & e band 100s ($1/M's)', value='680'},{type='cost', field='29 versions', value='250'}]

Here is the match:

{type=’cost’, field=’flag & e band 100s ($1/M’s)’, value=’680′},{type=’cost’, field=’29 versions’, value=’250′}

I understand why this string was matched. I do not understand why no other strings were matched. I expected the other matched strings to be:

1. {type=’cost’, field=’flag & e band 100s ($1/M’s)’, value=’680′},

2. {type=’cost’, field=’29 versions’, value=’250′}

Why were these matches not made?

Answer 1

The problem is using .* inside the subgroups. The type=\'(.*)\' matches greedily, i.e. it will yield cost', field='flag & e band 100s ($1/M's)', value='680'}, {type='cost.

Also: The delimiters in your data are also present in the content, e.g. your pattern seeks to parse field=\'(.*)\' but hits hard onto field='flag & e band 100s ($1/M's)', (note the extra 'after the M.

So I propose (if you also want to collect the content of the fields):

1. Watch the possible content of your “fields”, to shield the subgrouping (escape or find better delimiters, if data comes from external source).
2. Avoid greediness like Steve Wang mentions in his answer, to collect the subgroups only as intended.

otherwise, trigger only on the curly brace grouping, i.e. \{[^\}]+\}