Recently I have been doing some database searches for subscribers to our
list and, in light of something Eric said a month or so ago in response to
a previous question I had about db searches, I've been wondering about
efficiency in searches.
 
We're searching in a list which has 3 - 4 years of notebooks.  We know the
exact item wanted had "abracadabra" in it.  A couple of others had this
also, but the one we want has a unique subject line which will be
recognized when seen.  We know the month of the message, so we send:
 
// Job Echo=No
Database Search DD=Rules
//Rules DD *
s abracadabra in listname from 94/04/01 to 94/04/30
i
/*
 
We examine the returned index and see we want only item 019563. Now we get
to the question: how should the search be structured to minimize computer
time to do the search for the "print."
 
Leaving out the other material, here are variations:
 
s abracadabra in listname from 94/04/01 to 94/04/30
p 019563
 
same as above, but leave out the end date
 
s abracadabra in listname
p 019563
 
s * in listname
p 019563
 
From what Eric has said I have gotten the impression that the last might
be the most efficient, since we know the item number.  I sometimes get
faulty impressions though, so I'm asking.  Just something I've been
wondering about for a while.
 
Douglas
        Douglas Winship    Austin, Texas    [log in to unmask]
                    Secondary AUTOCAT Listowner