Hi Jeff,

The answer depends partly on whether you are using UniVerse or Unidata.

In general, a dimensioned matrix will be faster because you can index directly 
to the required item, however, you have to trade this
against the cost of doing MATREAD or some equivalent to populate it. Also 
inserting and deleting items is tough.

A dynamic array requires the system to search for the desired element which 
clearly could take a long time if you want something at
the far end. Inserting and deleting, however, is very quick.

The UniVerse/Unidata difference is that UniVerse maintains a "hint pointer" 
that points to the last field that you extracted. Thus,
if you walk linearly from one end to the other (in either direction) it can use 
the hint as a starting point and get the result very
quickly. If you hop around all over the place, the advantage of the hint is 
lessened but usually still good.

My understanding is that Unidata does not support field hints.

The best thing is to try experiments with realistic data sizes and access it in 
the same way your application would.


Martin Phillips
Ladybridge Systems Ltd
17b Coldstream Lane, Hardingstone, Northampton NN4 6DB, England
+44 (0)1604-709200


_______________________________________________
U2-Users mailing list
[email protected]
http://listserver.u2ug.org/mailman/listinfo/u2-users

Reply via email to