How can I match this array-like notation using regex in PHP?

Question

I'm trying to match the following array-like pattern with regex:

foo[bar][baz][bim]

I almost have it with the following regex:

~([^[]+)(?:\[(.+?)\])*~gm

However, the capturing groups only include:

Full match: foo[bar][baz][bim]
Group 1: foo
Group 2: bim

I can't figure out why it's only capturing the last occurrence of the [] structure. I'd like it capture foo, bar, baz, and bim in this case.

Any ideas on what I'm missing?

Answer 1

Repeated capturing groups in PCRE don't remember the values of each previous pattern. For this you need to invoke \G token:

(?|(\w+)|\G(?!\A)\[([^][]*)\])

See live demo here

Regex breakdown:

• (?| Start of a branch reset group
  • (\w+) Capture word characters
  • | Or
  • \G(?!\A) Conitnue from where previous match ends
  • \[ Match an opening bracket
  • ([^][]*) Capture any thing except [ and ]
  • \] Match a closing bracket
• ) End of cluster

PHP code:

preg_match_all('~(?|(\w+)|\G(?!\A)\[([^][]*)\])~', 'foo[bar][baz][bim]', $matches);
print_r($matches[1]);