From: Michael Fisher 
@Date: Fri, 07 Mar 2003 09:10:48 -0800
@Sender: semware-owner{at}sawasdi.apana.org.au



>From "Greg Macdonald" 
>Date Thu, 6 Mar 2003 17:18:17 -0800
>Define "in one line".
>With more complex searches, it occurs to me that the goal is 
>really more logical content centric than line centric and 
>the condition should be "in one sentence or paragraph".

In source code or data files, "in one line" usually means
"contained within one pair of line-delimiting control codes", 
such as , , or combinations thereof.

The line is an important structural and logical unit unit in 
such context.  Sentences and paragraphs would more apply to 
textual information in natural languages.

These two cconcepts require somewhat different approaches 
when you want to do complex searches.

>On the other hand, even a paper clip is inadequate if
>you have too many pieces of paper.

?  Can't follow this one.  
I'd rather say, there comes the point when even TSE and grep 
searches have to be replaced by something more specialized 
like awk.

>From "Charles Irish" 
>Date Thu, 6 Mar 2003 22:13:36 -0500
>Also if one doesn't like that, there is cludgy which would 
>be assign the strings to an array of strings and count the 
>hits by array occurence and if the count = the number of 
>strings in the array being tested, then you have a match on 
>all the strings 1..n and you can return in this case the 
>line in question

Yep, you could write a macro function for such searches, 
though it would have to be a rather complex one, to deal with 
all the eventualities of string searches.

If I understand you correctly, you want to search for the 
strings in question using an array, index the lines of the 
file by number of hits, and the lines that have as many hits 
as there are search strings must contain them all?




--
TSEPro Mailing List
List Maintenance: www.semware.com/html/list.htm



---
[sawasdi.apana.org.au] (3:800/846.13)