filter command and "not" wildcards

[Ben Rubinstein]

on 6/12/01 9:10 PM, Sivakatirswami at katir at hindu.org wrote:

>> filter <container> excluding <expression>
>> 
>> It should be one of those things where a relatively small programming effort
>> produces a considerable expansion in facility.
> 
> This is very fast:
> 
> repeat for each line x in dataTofilter
>   if x contains (expression) then
>      next repeat
>   else
>      put x & cr after dataOutPut
>   end if
> end repeat
> 
> put dataOutPut into url "file/processed/theFilteredData.txt"
> 
> I have processed lists up to 120,000 lines with the above in under 2 secs

A tribute to the speed of MC/Rev! but to have the same effect as the
proposed 'not filter', you'd have to replace the use of 'contains' with a
call to 'matchChunk'.  I don't know, but I'd guess the performance hit here
could be more substantial; often regular expression matches have to do quite
a bit of setup as they parse the expression, and in the absence of the 'not
filter', using a routine like that you propose above, this setup cost would
be paid 120,000 times in your example, rather than once.

I agree - this isn't a case in which we can't do what we want within the
current language, it's just something where what seems to be a small change
could provide much increased performance.  But if I had to choose, I'd
rather get features that would unlock things I currently do at all in
Rev/MC, like access to styled text in the clipboard, or better printing
facilities.

 
  Ben Rubinstein               |  Email: benr_mc at cogapp.com
  Cognitive Applications Ltd   |  Phone: +44 (0)1273-821600
  http://www.cogapp.com        |  Fax  : +44 (0)1273-728866