Need help with StringRegExpReplace

[DrJeseuss]

I feel what I'm trying to do is simple, though I can't even get a somewhat functional script going to do it... I have an xml file with tags enclosed as normal <tag> but in several places of several files I have some that are on multiple lines, such as:

<tag
attrib="0"
attrib="1"
>

I need to remove the CR and LF from this to give me:

<tag attrib="0" attrib="1">

I do NOT want to remove ALL CR and LF characters, only those within tags. I would expect what I want to do is find a pattern of <*> where * can be anything, then within * I need to replace any chr(10) or chr(13) with "". Now, saying isn't doing. Here's my latest NONWORKING try at it. I thought it would find < then anything then CR or LF, then anything then > and replace the same but without any CR or LF betweek the < and > Any Suggestions?

Local $read = FileRead("ReplaceTestIn.xml")
$Search = '[<](.*?)([\n|\r])(.*?)[>]'
$Replace = '<${1}${3}>'
$read = StringRegExpReplace($read, $Search,$Replace)
FileWrite("ReplaceTestOut.xml",$read)

[DrJeseuss]

I've found a very non-elegant but functional workaround for this problem... Even still, I'd be interested i hearing from the guru's about this so I can better understand how this works. Anyone?

[jchd]

When there is no match, there is no replacement so you get the input subject verbatim as result.

I don't have off-hand a clean one-liner solution for the general problem you submit but you may try something like

$pattern = "(?x) ([^<]*) (<) (.*) (s) (.*) (s) (.*) (s) (.*) (s) (.*) (s) (>) ([^<]*)"
$repl = "$1$2$3$5$7$9$11$13$14"

Note: the x option (?x) permits non-significant whitespaces in the pattern so that it's more readable.

I confess it's not clever but it will process the following input correctly:

Place random
text
bafore<tag
attrib="0"
attrib="1"
> some
text
after

[Malkey]

This method uses StringRegExpReplace to capture the text between "<tag" and ">", then uses a nested StringRegExpReplace to remove all CR and LF characters from the captured text only.

#cs
    ----- Contents of "TestTag.txt" file -------
    Place random
    text
    before<tag
    attrib="0"
    attrib="1"
    > some
    text
    after
#ce

$pattern = "(?s)(^.*<\s*tag)(.*?)(>.*$)"
$repl = "$1 "" & StringRegExpReplace('$2','(\\v+)', '') & ""$3"
;Space here^, could change to @LF, and no space here ^ between attributes, could change to a space (referencing the line above).

;ConsoleWrite('"' & StringRegExpReplace(FileRead("TestTag.txt"), $pattern, $repl) & '"' & @LF)
ConsoleWrite(Execute('"' & StringRegExpReplace(FileRead("TestTag.txt"), $pattern, $repl) & '"') & @LF)

#cs
    Returns:-
    Place random
    text
    bafore<tag attrib="0"attrib="1"> some
    text
    after
#ce

[jchd]

Hmm, I'd call that a cheating one-liner!

[hawkair]

After spending nearly 30 minutes to understand how the "nesting" example works

I believe that ,at least, in the way it is written it does only one substitution (one tag)

I added one more tag to the testtag file and I got nothing

I would do it like this

expandcollapse popup

#cs
Contents of testtag.txt
<td
valign="top"
align="Left">
<b><font
size="2"><font
color="white"
>Greek
</font></font></b>
style="float: left;
padding-top: 2px;"></span> <input
value="Display"
style="margin: 0px;
padding: 0px;
width: 80px;
font-size: 10px;"
type="button">
    Place random
    text
    before<tag
    attrib="0"
    attrib="1"
    > some
    text
    after
#ce

$txt = FileRead ("testtag.txt")
$tag = StringRegExp ($txt, "(?s)<.*?>", 3)
For $i = 0 To UBound ($tag) - 1
    $txt = StringReplace ( $txt, $tag[$i], StringRegExpReplace ($tag [$i], " *rn *", " "))
Next
MsgBox (0, "", $txt)
Exit

#cs
    Returns:-
<td valign="top" align="Left">
<b><font size="2"><font color="white" >Greek
</font></font></b>
style="float: left;
padding-top: 2px;"></span> <input value="Display" style="margin: 0px; padding: 0px; width: 80px; font-size: 10px;" type="button">
    Place random
    text
    before<tag attrib="0" attrib="1" > some
    text
    after
#ce

It does leave one blank before the ">" if it is on the start of the line, but maybe somebody else can fix this

Edited March 18, 2012 by hawkair

[Malkey]

@hawhair

Using your test data and the essence of your example script, this example works.

$tag = "'" & StringRegExpReplace(FileRead("testtag.txt"), "(?s)(<.*?>)", ''' &  StringRegExpReplace (''$1'',  " *v+ *", " ") & ''') & "'"
ConsoleWrite(Execute($tag) & @LF)
MsgBox(0, 0, Execute($tag))

[hawkair]

@Malkey

Indeed it does!!

It is amazing!!

thank you very much

I will keep on studying your script so that maybe some day I will be able to create something similar