changing the first occurrence only of a string in a line

[Bernard Devlin]

Peter, given the sample you provide, I believe this will work:

on mouseUp
   put "A aa TAB B bb TAB B bb TAB B bb TAB C cc TAB D dd TAB D dd" &
cr after tData
   put "A aa TAB B bb TAB B bb TAB B bb TAB C cc TAB D dd TAB D dd" &
cr after tData
   --replace "TAB" with tab in tData
   put "B 1!" & cr after tSubs
   put "C 2!" & cr after tSubs
   set the wholematches to true
   repeat for each line tLine in tSubs
      put word 1 of tLine into tFind
      put word 2 of tLine into tSubstitute
      put wordoffset(tFind,tData) into tFoundPos
      if tFoundPos > 0 then
         put tSubstitute into word tFoundPos of tData
      end if
   end repeat
   put tData
end mouseUp

However, one of the frustrating things about Rev is that the PCRE
engine behind replaceText and matchText is so poorly documented.  For
example, replaceText can have its functionality altered by the
inclusion of modifiers such as "(?i)" for case insensitive and "(?m)"
for multiline.  There are other modifiers
(http://uk.php.net/manual/en/reference.pcre.pattern.modifiers.php)
some of which are unacceptable, and others that seem to be acceptable.
 Nowhere have I found these documented with regard to Rev.  The (?A)
anchored modifier (if it was one of the acceptable tokens) may have
worked in your case, but it is apparently unnaceptable.

Really you ought to just be able to take your SED expressions and use
them with replaceText, as SED expressions are regex AFAIK.  There may
be some differences between SED and pcre, but since the differences
between pcre and Rev are not documented it is vexatious.

Bernard

On Fri, Jun 19, 2009 at 7:33 AM, Peter
Alcibiades<palcibiades-first at yahoo.co.uk> wrote:
> How would you do the following in Rev?
>
> We have a file consisting of records with tab separated fields.  Each field
> has a tag followed by contents.  Some tags occur more than once in some
> records which thus have varying numbers of fields. Duplicates are always
> consecutive.  I want to eliminate all the occurrences of any tag except
> the first one.  The duplicate tags can occur any place in the record, but
> if they are duplicated, will always be consecutive.
>
> Doing this in SED is not particularly difficult, but it does require going
> out to shell, and so its not cross platform.  You just change the tag
> using the local scope to something else.  SED then only changes the first
> occurrence in a record.  Then you use the global scope and change all of
> them.  Then you go back and change the first one back to what it was.  In
> fact, if using SED like this, the only thing you need it for is to do the
> local, first tag, change - once this is done, the rest can be done in Rev.
> But it would be nice to stay in Rev for the whole thing.
>
> Is there a way in Rev to pick the first occurence of a string in a record,
> change it and not subsequent occurences, and then move on to the next
> record and do the same thing?
>
> That is, mimic the 'local' editing mode of SED?
>
> Bet you all thought them dinosaurs like SED had to be extinct by now!  But
> no, they are still trampling around in the swamps of text manipulation....
>
> For the sake of clarity, a record might look like this:
>
> A aa TAB B bb TAB B bb TAB B bb TAB C cc TAB D dd TAB D dd
>
> and what is wanted is to change the first occurrence of B to, for
> example !1, and the first occurence of D to, for example !2, or anyway
> something that will not occur by chance, to allow the subsequent editing
> to work globally on the file.  This is what SED does in local mode.
>
> Peter
> _______________________________________________
> use-revolution mailing list
> use-revolution at lists.runrev.com
> Please visit this url to subscribe, unsubscribe and manage your subscription preferences:
> http://lists.runrev.com/mailman/listinfo/use-revolution
>