Regex: how to exclude a word from the found string

[leuce]

Hello everyone

I'm trying to perform some regex find/replace on an XML file in which only some tag pairs must be affected, and the clue to know which tag pairs must be affected comes *after* the tag pair.

Here's a simplified example of my text:

<target>some text</target> some text <target>some text </target> some text <target>some text</target> <reference> some text <target> some text </target>some text

I want to replace "<target>some text</target> <reference>" with "<target>{{some text}}</target> <reference>".  But I can't figure out how to tell the regex not to include all the other "target"s that come before the text that I really want to match.

My limited regex knowledge tells me that this ought to work:

"(<target>)(.+?)^(<target>)(</target> <reference>)", "$1{{$2}}$4"

...but it doesn't.  If I omit the ^(<target>), then AutoIt matches everything from the first <target> all the way up to the <reference>.

Can this be done?

Thanks

Samuel

 

[leuce]

By the way, I solved my problem by using StringReverse:

#include <MsgBoxConstants.au3>
$foo = "<target>some text</target> some text <target>some text </target> some text <target>some text</target> <reference> some text <target> some text </target>some text"
$foofoo = StringReverse ($foo)
$bar = StringRegExpReplace ($foofoo, "(>ecnerefer< >tegrat/<)(txet emos)(>tegrat<)", "$1}}$2{{$3")
$barbar = StringReverse ($bar)
MsgBox (0, "", $barbar, 0)

...but I'm thinking that this must be possible using regex instead of a kludge.

[iamtheky]

 

$foo = "<target>some text</target> some text <target>some text </target> some text <target>some text</target> <reference> some text <target> some text </target>some text"
msgbox(0, '' , stringreplace($foo , "<target>some text</target> <reference>" , "<target>{{some text}}</target> <reference>"))

 

Edited June 27, 2016 by iamtheky

[leuce]

  On 6/27/2016 at 1:17 PM, iamtheky said:

 

$foo = "<target>some text</target> some text <target>some text </target> some text <target>some text</target> <reference> some text <target> some text </target>some text"
msgbox(0, '' , stringreplace($foo , "<target>some text</target> <reference>" , "<target>{{some text}}</target> <reference>"))

 

Expand  

Thanks, but "some text" in my example is a placeholder for some text.  I don't know in advance what that text is going to be.

[Melba23]

leuce,

You need to make the RegEx non-greedy - then it looks for the shortest match:

$sText = "<target>some text</target> some text <target>some text </target> some text <target>some text</target> <reference> some text <target> some text </target>some text"

$sNewText = StringRegExpReplace($sText, "(?U)(<target>)(some text)(</target> <reference>)", "$1{{$2}}$3")

ConsoleWrite($sNewText & @CRLF)

M23

[leuce]

  On 6/27/2016 at 1:28 PM, Melba23 said:

leuce,

You need to make the RegEx non-greedy - then it looks for the shortest match:

$sText = "<target>some text</target> some text <target>some text </target> some text <target>some text</target> <reference> some text <target> some text </target>some text"

$sNewText = StringRegExpReplace($sText, "(?U)(<target>)(some text)(</target> <reference>)", "$1{{$2}}$3")

ConsoleWrite($sNewText & @CRLF)

M23

Expand  

Thanks, Melba23, but "some text" in my example is just some text.  So I'd have to update your suggestion with something like .+?, i.e. "anything".  But this:

"(?U)(<target>)(.+)(</target> <reference>)", "$1{{$2}}$3")

yields this:

<target>{{some text</target> some text <target>some text </target> some text <target>some text}}</target> <reference> some text <target> some text </target>some text

instead of this:

<target>some text</target> some text <target>some text </target> some text <target>{{some text}}</target> <reference> some text <target> some text </target>some text

 

[Melba23]

leuce,

I see the problem, but I am afraid I have no idea how to solve it. We will need to wait for one of the gurus to pass by.

M23

[Melba23]

leuce,

Having just said that, this seems to work:

$sText = "<target> some text </target> some text <target> some text </target> some text <target>some text</target> <reference> some text <target> some text </target>some text"

$sNewText = StringRegExpReplace($sText, "(?U)(<target>)([^<]*)(</target> <reference>)", "$1{{$2}}$3")

ConsoleWrite($stext & @CRLF & $sNewText & @CRLF)

The RegEx is asked not to include any "<" characters in the intervening text - which means that only the actual text we are looking for will work (unless you have any of them in the text itself).

M23