Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!elroy.jpl.nasa.gov!ncar!midway!gsbacd.uchicago.edu!gft_robert
From: gft_robert@gsbacd.uchicago.edu (opcode ranger)
Newsgroups: comp.sys.mac.hypercard
Subject: Need advice on speeding up card find in mega-sized stack
Message-ID: <1991May15.114727.1@gsbacd.uchicago.edu>
Date: 15 May 91 17:47:27 GMT
Sender: news@midway.uchicago.edu (NewsMistress)
Lines: 37

I have a big problem (literally).  I've set up a stack for my clients which
is about 12 MB in size (yes, they've got a LOT of data: about 12k cards each
containing 13 pieces of info, in addition to fields, buttons, etc).

Problem is, it's slow.  Let's say that the stack is keeping track of years,
with each year  having it's own card (this isn't the situation, but just
an example).  OK, let's say a person wants to go to 1991.  My current system
uses this method: each card is named something like "Year 1932", or something,
i.e. "Year" && the year itself.  Thus if a person wants to go to 1991, I do
something like this:

on goToYear theYear
	put "Year" && theYear into theYear
	go to card theYear
end goToYear

This is VERY slow, taking as much as a minute.

What's the best way to do this.  Would go to cd id be quicker (requiring me
to keep a list of cd id's somewhere else)?  Would keeping the cards sorted
and doing a binary search be the best way?

In general, is HC a good medium for handling so much data (the stack will
be used for large data updates, requiring the (automatic) updating
of perhaps a couple of hundred cards per session)?  


Any info much appreciated!

Robert

-- 
============================================================================
= gft_robert@gsbacd.uchicago.edu * generic disclaimer: * "Good tea.        =
=            		         * all my opinions are *  Nice house."     =
=                                * mine                *  -Worf            =
============================================================================