How to extract multiple patterns from a multi-line string

Question

I have a string that looks like this. It's obviously a multi-line string and I would like to split it into one string per stanza.

{
   "timestamp":1317911700,
   "application":"system.dev",
   "metrics":{
      "qlen":0,
      "read.bytes":0,
      "write.bytes":185165.0123762,
      "busy":0.021423
   },
   "dimensions":{
      "device":"sda"
   }
}

{
   "timestamp":1317911700,
   "application":"system.fs",
   "metrics":{
      "inodes.used":246627,
      "inodes.free":28703901,
      "capacity.kb":227927024,
      "available.kb":209528472,
      "used.kb":6820512
   },
   "dimensions":{
      "filesystem":"/"
   }
}

{
   "status_code":0,
   "application":"system",
   "status_msg":"Data collected successfully"
}

My regex looks like this:

/^({\n[^}]+^})/m

But I am only capturing:

{
   "status_code":0,
   "application":"system",
   "status_msg":"Data collected successfully"
}

Which kinda makes sense since that's where the first curly brace is. What I am trying to do is capture from anywhere there is a /^{/ to anywhere there is a /^}/ as a single string. But I think the other curly braces in there are tr

Is there a reason you can't use a real JSON parser to extract the data you want? — AFresh1
– AFresh1, Commented Oct 6, 2011 at 17:57
It's not valid JSON. It would have to look like this to be valid. [ {stanza1},{stanza2},{stanza99}]. I've already run it through a validator and it failed as is. I do not control the output so I need to capture it and munch it myself. I was wondering if I could do something with lookahead. my @foo = /somelookahead/; — gdanko
– gdanko, Commented Oct 6, 2011 at 18:02
Nevermind, I see that there are no commas at the end of each stanza, so JSON parsing won't work. — AFresh1
– AFresh1, Commented Oct 6, 2011 at 18:06

ikegami · Accepted Answer · 2011-10-06 18:11:33Z

4

I can think of a few approaches.

There is an example somewhere in perlre on how you can implement a recursive pattern. This is hard. You need to take curlies in strings into account.
Text::Balanced already provides means of matching balanced parens (including curlies). This might be easier, because I think it can take curlies in strings into account.
It looks like you can simply split on blank lines.
```
@json_snippets = split /^$/m, $json_snippets;
```
But the most reliable solution is to use JSON::XS's "incremental parser". (Search for that in its documentation.)

edited Oct 6, 2011 at 18:11

answered Oct 6, 2011 at 18:04

ikegami

391k17 gold badges291 silver badges555 bronze badges

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

Comments

Alex Howansky · Accepted Answer · 2011-10-06 18:06:12Z

1

for my $stanza (split /^$/m, $str) {
  ...
}

answered Oct 6, 2011 at 18:06

Alex Howansky

54k8 gold badges85 silver badges102 bronze badges

1 Comment

Kenny Over a year ago

And thereafter you have valid JSON strings: foreach my $stanza ( split( /^$/m, $str ) ) { my $json = decode_json( $stanza ); print Dumper( $json ); }

AFresh1 · Accepted Answer · 2011-10-06 18:09:47Z

0

If you can't use a JSON parser to properly do it, I would just split at the end of a stanza.

my @stanzas = split /^}\K\n\n/;

edited Oct 6, 2011 at 18:09

answered Oct 6, 2011 at 18:03

AFresh1

4062 silver badges5 bronze badges

1 Comment

ikegami Over a year ago

It's not valid JSON. It's more like a series of JSON snippets.

Collectives™ on Stack Overflow

How to extract multiple patterns from a multi-line string

3 Answers 3

Comments

1 Comment

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

1 Comment

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related