Subject trying to avoid large datasets
Author markd_mms
i'm in the process of porting a music catalogue maintenance program
from clarion (which natively handles large datasets well) to c++.

entering new title into database revolves around a screen that lists
*all* the titles in the catalogue (about 120000) which takes 10
seconds or so to load in order (unordered takes almost no time). the
data entry operators like this because it allows them to locate where
an item might be and then quickly scroll up and down to see if it's
already been entered but with a spelling mistake.

i realise this is more of ye olde paradox way of doing things but
they've said it's a compulsory feature so there's not much i can do. i
was wondering if i could fudge some similar kind of functionality by
doing something like...

SELECT FIRST 10 ...
FROM ...
WHERE param = 'whatever they enter'
ORDER BY title

but also selecting items preceeding whatever they're searching for?
kinda like

SELECT PREVIOUS 10 ...
FROM ...
WHERE param = 'whatever they enter'
ORDER BY title

so it might return

whatever they eeter
whatever they enter
whatever they etner

TIA