Hello, I need help on matching and replacing text with different matched text from same string. My goal is to be able to batch process a large set of text files that each have similar patterns and replacement criteria. Below is an example of a typical case. In the example below, BOLD text replaces strikeout text:

TITLE: Beautiful Losers - 8/22/2004 12:43:00 AM
AUTHOR: mt San Francisco
DATE: 8/22/2004 12:43:00 AM
-----
BODY:
Yerba Buena Center for the Arts
Through October 10, 2004

Reviewed by Clark Buckner


....  text of varying length here ...


<!-- BIO>Clark Buckner works as the gallery director at Mission 17 and writes the weekly "Critic's Choice: Art" column for the <i>San Francisco Bay Guardian</i>.*review by Clark Buckner~--><!-- MENU>Beautiful Losers by Clark Buckner~-->
-----
--------

AUTHOR: mt, San Francisco

 replace

mt, San Francisco

 with:

Clark Buckner

to get:

AUTHOR: Clark Buckner

<!-- BIO>Clark Buckner works as the gallery director at Mission 17 and writes the weekly "Critic's Choice: Art" column for the <i>San Francisco Bay Guardian</i>.*review by Clark Buckner~--><!-- MENU>Beautiful Losers by Clark Buckner~-->

I've experimented on regex simulator tools and am just hitting a wall on how to form the expression. I'm working in an environment that supports standard perl regex. Any help would be very much appreciated. Many thanks.

Try a matching pattern of:

(AUTHOR:\s*)([^\r\n]*)([\S\s]+)\*review\s+by\s+([^~]*)

and replacement text of

$1$4$3$4

with the 'ignore case' option set (or the text altered to suit). You don't mention the regex/language tht you are using, to the replaecment text may need to be "\1\4\3\4" or whatever is appropriate.

Susan

PS: Only tested with the single example you provided and making assumptions about there never being a "*review by" within the text before the one you are after.

Aussie Susan:

Try a matching pattern of:

(AUTHOR:\s*)([^\r\n]*)([\S\s]+)\*review\s+by\s+([^~]*)

and replacement text of

$1$4$3$4

with the 'ignore case' option set (or the text altered to suit). You don't mention the regex/language tht you are using, to the replaecment text may need to be "\1\4\3\4" or whatever is appropriate.

Susan

PS: Only tested with the single example you provided and making assumptions about there never being a "*review by" within the text before the one you are after.

Hi Susan, thank you for the quick reply. This is very close to exactly what I need, just a couple tweaks and I think we'll have it.

I'm working with a text processor which supports the PCRE library so the "\1\4\3\4" replace format was the one that worked.

Your assumption that "*review by" never occurs within the text before the one I'm after is partially correct. My fault for not being more clear:

"*review by" isn't a constant, what I meant is "*(some wildcard text) by "

i.e. "*by", "*review by", "*a review by", "*an essay by", etc. would all be valid delimiters. If it helps, another thing to note is that this match will always be from within the block of text that begins with <!-- BIO>.

update:

After some experimentation, I got it working by using

*[\S\s]*?\s*?by

instead of

*review\s+by

like this:

(AUTHOR:\s*)([^\r\n]*)([\S\s]+)\*[\S\s]*?\s*?by\s+([^~]*)

So now the only tweak is after running the expression, the string "*review by" is getting deleted in the result. This text needs to remain unaffected. Any thoughts?

Thanks again!

best,

--dhl

dhl,

Add another set of parentheses around the part of the expression that is matching "*review by":

(AUTHOR:\s*)([^\r\n]*)([\S\s]+)(\*[\S\s]*?\s*?by\s+)([^~]*)

Modify your replacement to use the new group:

\1\5\3\4\5

That should do it.

Jeff

jeff.hillman:

dhl,

Add another set of parentheses around the part of the expression that is matching "*review by":

(AUTHOR:\s*)([^\r\n]*)([\S\s]+)(\*[\S\s]*?\s*?by\s+)([^~]*)

Modify your replacement to use the new group:

\1\5\3\4\5

That should do it.

 

Jeff

Jeff,

Perfect! Thank you very much!

best,

--dhl