Re: Array vs Temporary Tables, benchmark testing

[Krisjanis Gale <Krisjanis.Gale@NY.FRB.ORG>]

actually, you'd *think* that arrays would be the way to go...
but often it's more difficult to sit down and invent SQR
arrays that behave like a DB table, with roughly the
same structure and variable/column names,
that it's just more convenient to rely on your DB
more directly.

performance gains i've gotten using arrays
have been pretty small, at least on our Oracle7 platform
(which is very distributed over a network).

think of arrays as a better way of removing yourself
from the DB entirely.  if what you want to do is quickly
extract a ton of information, without thinking in terms
of SQL, but rather, program logic (SQR), then arrays
*are* better.

but if you're performing lots of DB transactions with
the data you're gathering, temporary tables are
much better.

long story short,
it depends on what you're doing =)


(kris)janis p. gale
hrsd - federal reserve bank of new york
x8163

>>> Jason Wainwright <JasonW@LIMON.CO.ZA> Tuesday, September 14, 1999 2:42:45 AM >>>
Hi All

There are many routines that need data to be stored and/or sorted.  I was
wondering whether, performance wize, if arrays are better than temporary
tables.  I am led to believe that some DBMS' will only create indexes if
many rows are being used (e.g. Sybase may have 1000).  And so with all the
overhead ,such as server and network resources,  how expensive are arrays,
lookups and temporary tables in sqr?  And how can I measure performance?
And what performance issues are there to consider?

THanx