Hello to all:

What are the pros and cons of defining data as numeric (with a defined
precision) vs. double?  I am pulling data from our corporate DB2 database
and many of the fields are defined as numeric.  I typically prefer using a
double data type.  Is it 6 of 1 and half a dozen of the other?

Incidentally, there was a limited discussion some time ago about pulling
data into R:Base through an IBM DB2 ODBC client connection.  Results would
come back quickly when using MS-Access but very slowly using R:Base.  I have
since come to a new understanding that when a select clause is submitted to
DB2, since it's a true client/server DBMS, you have to wait until the server
is done gathering the data before you start to see results.  I've been used
to getting near instant results when using R:Base.  Additionally, DB2 can be
specifically optimized for use with MS-Access and Visual Basic.  I wish I
knew how that was done so that something similar could done with an R:Base
connection.  At any rate I am now downloading data from the DB2 server to
the tune of 2.5 million records in about an hour and a half.  Not too bad
for still getting to use my trusty R:Base.  You just have to be careful
about how you define the select statement when querying the DB2 database.

Thanks for any and all advice on the data type question.

Mike Ramsour
R:Base database administrator
c/o AK Steel Coshocton Works
Quality Department

Phone:  740-829-4340

Reply via email to