regex: splitting conditionally on |

[computorist]

I'm parsing mediawiki markup and I'd like to split a multi-line string
on | (vertical-bar), but only if it isn't contained w/in another
pattern.

| name1 = value1
| name2 = value2 | name3 = value3
| name4 = [[foo|bar]]

I'd like String.split(/pattern/) to return ['name1 = value1','name2 =
value2','name3 = value3','name4 = [[foo|bar]]']

I've working on a negative look-ahead pattern without success and my
brain is tired.

Thanks for any suggestions

 

[Robert Klemme]

I'm parsing mediawiki markup and I'd like to split a multi-line string
on | (vertical-bar), but only if it isn't contained w/in another
pattern.

| name1 = value1
| name2 = value2 | name3 = value3
| name4 = [[foo|bar]]

I'd like String.split(/pattern/) to return ['name1 = value1','name2 =
value2','name3 = value3','name4 = [[foo|bar]]']

I've working on a negative look-ahead pattern without success and my
brain is tired.

Thanks for any suggestions

Without testing, something like this might work:

str.scan %r{
(?: \[ [^\]]* \] | [^|] )+
}xm

Cheers

robert

 

[Sebastian Hungerecker]

I'm parsing mediawiki markup and I'd like to split a multi-line string
on | (vertical-bar), but only if it isn't contained w/in another
pattern.

| name1 = value1
| name2 = value2 | name3 = value3
| name4 = [[foo|bar]]

I'd like String.split(/pattern/) to return ['name1 = value1','name2 =
value2','name3 = value3','name4 = [[foo|bar]]']

I've working on a negative look-ahead pattern without success and my
brain is tired.

Thanks for any suggestions

str.split /\s*\|\s*(?=\s*\w+\s*=)/
I don't know whether this exactly meets your requirement, but this will split
only on |s that are followed by a word and a =. For your sample input that
gives the desired result.

HTH,
Sebastian

 

[computorist]

I'm parsing mediawiki markup and I'd like to split a multi-line string
on | (vertical-bar), but only if it isn't contained w/in another
pattern.

| name1 = value1
| name2 = value2 | name3 = value3
| name4 = [[foo|bar]]

I'd like String.split(/pattern/) to return ['name1 = value1','name2 =
value2','name3 = value3','name4 = [[foo|bar]]']

I've working on a negative look-ahead pattern without success and my
brain is tired.

Thanks for any suggestions

str.split /\s*\|\s*(?=\s*\w+\s*=)/
I don't know whether this exactly meets your requirement, but this will split
only on |s that are followed by a word and a =. For your sample input that
gives the desired result.

HTH,
Sebastian

This seems to work well. Searching for what I want is better than
excluding the bits I don't want.

Thanks