RegEx Question for scripted HTML code modification

[rudi]

Hi.

For a MRTG setup I want to quickly extract all PortChannels to a separate HTML file.

The RegEx matching the links I want to preserve is this: (?i)(?s)(<td><DIV><B>Po.*?</td>)

All other <td>....</td> are to be removed.

I was wondering about nesting a 2nd match for <td>...</td> in a different "bracket level", so that the one I want to preserve will be $1, the other one $2, and then just to "replace $1", so that all the $2 will be dropped.

This must be easy, I just can't see it

$index="\\nw04goe\sys\apache2\htdocs\mrtg\index.html"
$PO_Only="\\nw04goe\sys\apache2\htdocs\mrtg\index-PO.html"
Dim $aIndex

$Content=FileRead($index)
FileWrite($PO_Only,StringRegExpReplace($Content,"?????????????????","$?$?")

Regards, Rudi.

Edited December 3, 2010 by rudi

[jchd]

It isn't clear to me if this is "easy".

Rewriting your conditions we have that you want to replace any <td>....</td> which doesn't contain a match to (?i)(?s)(<td><DIV><B>Po.*?</td>) by an empty string.

But what do you do with the rest?

If it was me, I'd extract the matching TDs using this pattern with StringRegExp in mode 3 and I would insert the captures (array) into a clean homebrew html squeleton. That would keep you safe if ever the structure of the input html changes, as long as the valuable TDs continue to match the simple pattern you use. You also are free to manage the output html as you wish.

[rudi]

Hi.

Thanks for your reply.

I was wondering about using stringregex(), mode 3 as well, but I would loose all the other HTML content.

This is, how it works really nice:

$index = "\\nw04goe\sys\apache2\htdocs\mrtg\index.html"
$PO_Only = "\\nw04goe\sys\apache2\htdocs\mrtg\index-PO.html"
Dim $aIndex



$Content = FileRead($index)
$Content = StringRegExpReplace($Content, "(?i)(?s)((<td><DIV><B>Po.*?</td>)|<td><DIV><B>.*?</td>)", "$2")
FileWrite($PO_Only,StringRegExpReplace($Content,"(?i)(?s)(<tr>\s*</tr>)","")) ; remove empty table rows: <tr> 0 to unlimited whitespace </tr>

ShellExecute($PO_Only)

As I remember, how hard RegEx was in the beginning, I'll spend a few minutes, to explain it to beginners, that might come across this example. I remember very well, that most RegEx examples I came across were not readable to me.

(?i)(?s)((<td><DIV><B>Po.*?</td>)|<td><DIV><B>.*?</td>)", "$2"
(?i) = case insensitive
    (?s) Dot (".") does match new lines (the match doesn't care about line breakes in the HTML code)

        (                        |                    ) = the whole match goes to backreference 1, which would be addressed using "$1" (in stead of "$2") in the replace statement.
        (                        |<td><DIV><B>.*?</td>) = all the "Rule 2 matches" go to BackRef 1. Rule 2 is only checked, when Rule 1 *DIDN'T* match, so it doesn't matter
                                                          that Rule 2 also covers all stuff matched by Rule 1: If Rule 1 is matched, Rule 2 won't be checked.
        ( <td><DIV><B>Po.*?</td> |                    ) = that way Rule 1 would go to BackRef1 as well, so no way to distinguish.
        ((<td><DIV><B>Po.*?</td>)|                    ) = the extra brackets make "Rule 1 matches" to go to BackRef 2, $2 in the replace statement. Distinguish done! (nested grouping)
                                                          Note: Rule 1 matches *ALSO* go to BackRef 1, as the hole match goes to BackRef 1. (Just leave it unused :-))

.*? = match any character("."), as *LITTLE* times as needed("*?"). (Lazy). Using "*" (and not "*?") it would be "greedy" (match-as-much-as-possible), 
so it would match everything from the first opening "<td>..." to the last closing "...</td>" in the *WHOLE* HTML code

Regards, Rudi.

[edit: Typos]

Edited December 3, 2010 by rudi