RE: SEARCH all problems

It has been a while since I did any comparisons. For what I was using it
for, I'm pretty sure I remember that a binary search against a table was
faster than a random read to the corresponding data file. Of course, there
are lots of variables: length of table, size of data record, available
memory, etc.

If the data file can be updated while your program is running, you'll have
fun. Using a table would not show the change while a random read to the data
file would. Also, if the number of records in the file exceed the size of
your table, the program would have to be changed and recompiled.

I would recommend putting together a little benchmark program and see what
works best for your situation!

Have fun!
Richard



-----Original Message-----
From: cobol400-l-bounces@xxxxxxxxxxxx
[mailto:cobol400-l-bounces@xxxxxxxxxxxx]On Behalf Of Daniel Mielke
Sent: Monday, November 28, 2005 1:34 PM
To: 'COBOL Programming on the iSeries/AS400'
Subject: RE: [COBOL400-L] SEARCH all problems



Now that I have the SEARCH ALL thing down, I'm wondering; since the search
is being done in memory, which is faster, the binary search via memory, or
the DASD start/read action?