StringRegExpSplit

[genius257 <hot415@…>]

A core function to split strings by regular expression.

[Jos]

as in StringRegExp() you mean?

Jos

[genius257 <hot415@…>]

Yes and no.
Yes that it's by using regular expression.
No because StringRegExp returns the part of string matched by the pattern.

For example StringRegExpSplit("abc", "(:?)") would return ["a", "b", "c"]

[Melba23]

You mean like this:

#include <Array.au3>
$aRet = StringRegExp("abc", ".", 3)
_ArrayDisplay($aRet, "", Default, 8)

M23

[genius257 <hot415@…>]

In a small, simple example, yes.

StringRegExpSplit("|a

|b

c|

d|", "\s*\|\|\s*")

[ "|a", "|b", "c|", "d|" ]
Or

StringRegExpSplit("|a[ "|a", "

|b	c|	d|", "\s*(\|\|)\s*")
", "|b", "	", "c|", "	", "d|" ]

Or
StringRegExpSplit("Hello 1 word. Sentence number 2.", "(\d)")
[ "Hello ", "1", " word. Sentence number ", "2", "." ]

[genius257 <hot415@…>]

Well it seems i forgot the code block

Here's the examples with the code block

StringRegExpSplit("|a|| |b|| c| ||d|", "\s*\|\|\s*")
[ "|a", "|b", "c|", "d|" ]

StringRegExpSplit("|a|| |b|| c| ||d|", "\s*(\|\|)\s*")
[ "|a", "||", "|b", "||", "c|", "||", "d|" ]

StringRegExpSplit("|a|| |b|| c| ||d|", "(\d)")
[ "Hello ", "1", " word. Sentence number ", "2", "." ]

[Melba23]

The first 2 are easy to do with the existing function:

#include <Array.au3>

;StringRegExpSplit("|a|| |b|| c| ||d|", "\s*\|\|\s*")
;[ "|a", "|b", "c|", "d|" ]

$aRet = StringRegExp("|a|| |b|| c| ||d|", "(..)(?:\s?\|\||\z)", 3)
_ArrayDisplay($aRet, "", Default, 8)

;StringRegExpSplit("|a|| |b|| c| ||d|", "\s*(\|\|)\s*")
;[ "|a", "||", "|b", "||", "c|", "||", "d|" ]

$aRet = StringRegExp(StringStripWS("|a|| |b|| c| ||d|", 8), "(..)(\|\||\z)", 3)
_ArrayDisplay($aRet, "", Default, 8)

No idea about the last one as the given input does not match the required output in any way.

M23

[genius257 <hot415@…>]

Yeah, the last one was meant to be

StringRegExpSplit("Hello 1 word. Sentence number 2.", "(\d)")
[ "Hello ", "1", " word. Sentence number ", "2", "." ]

And i do get that StringRegExp CAN solve this, it simply seems like a ease of use function.
Especially when converting from one language (like js or php) to AutoIt

[Jos]

Something like this?

#include <Array.au3>

$aRet = StringRegExpSplit("|a|| |b|| c| ||d|", "\s*\|\|\s*")
_ArrayDisplay($aRet, "", Default, 8)
;[ "|a", "|b", "c|", "d|" ]

$aRet = StringRegExpSplit("Hello 1 word. Sentence number 2.", "(\d)")
;~ [ "Hello ", "1", " word. Sentence number ", "2", "." ]
_ArrayDisplay($aRet, "", Default, 8)

$aRet = StringRegExpSplit("|a|| |b|| c| ||d|", "\s*(\|\|)\s*")
;[ "|a", "||", "|b", "||", "c|", "||", "d|" ]
_ArrayDisplay($aRet, "", Default, 8)

func StringRegExpSplit($input,$regex)
	$a = StringRegExp($input, "(.*?)"& $regex & "|[\S]+", 3)
	return $a
EndFunc

Jos

[genius257]

Hmmm seems to work for almost everything, except the following.

#include <Array.au3>

$aRet = StringRegExpSplit("abc", "(?:)")
;expected [ "a", "b", "c" ]
;actual   [ "", "a", "", "b", "", "c", "" ]
_ArrayDisplay($aRet, "", Default, 8)

func StringRegExpSplit($input,$regex)
	$a = StringRegExp($input, "(.*?)"& $regex & "|[\S]+", 3)
	return $a
EndFunc

But the chances for that to give me problems currently are slim :)

[jchd18]

It can still work:

$aRet = StringRegExpSplit("|a|| |b|| c| ||d|", "\s*\|\|\s*")
_ArrayDisplay($aRet, "", Default, 8)
;[ "|a", "|b", "c|", "d|" ]

$aRet = StringRegExpSplit("Hello 1 word. Sentence number 2.", "(\d)")
;~ [ "Hello ", "1", " word. Sentence number ", "2", "." ]
_ArrayDisplay($aRet, "", Default, 8)

$aRet = StringRegExpSplit("|a|| |b|| c| ||d|", "\s*(\|\|)\s*")
;[ "|a", "||", "|b", "||", "c|", "||", "d|" ]
_ArrayDisplay($aRet, "", Default, 8)

$aRet = StringRegExpSplit("abc", "")
;[ "a", "b", "c" ]
_ArrayDisplay($aRet, "", Default, 8)

Func StringRegExpSplit($input,$regex)
	Return StringRegExp($input, "(.+?)"& $regex & "|[\S]+", 3)
EndFunc

But maybe you'll tell us that if the subject has adjacent separators it won't again work the way you expect it to.

Note that since the function is an actual one-liner, there is no real purpose for a new function: it's simpler to code it inline with the ad-hoc pattern suited to the case being handled.
Also Trac isn't the place to discuss various regexes; better post a topic in help and when & if a new function emerges from discussion then it'll be time to place a feature request here.