 |
 |
Archives of the TeradataForum
Message Posted: Sat, 30 Nov 2002 @ 16:49:24 GMT
| Subj: | | Re: Oracle vs Teradata compare |
| |
| From: | | Judge, James A |
Couple of thoughts. Assuming the Teradata schema to hold the Oracle schema is the same. If so, and you have most of your Oracle
columns defined as VARCHAR2 you should try to convert the columns to CHAR where appropriate in Teradata. This data discovery will
allow you to determine any COMPRESS values for columns. IF there are the usual number of default values in the CHAR columns, as well
as dates, and numeric then you can get significant dasd savings in Teradata. With minimal COMPRESS usage conservatively, I would say
15%; but the last two migration efforts I have been involved with have yielded 30-35% reduction in row size due to more extensive
evaluation and use of COMPRESS.
As far as indexing overhead it depends on expected access paths of the data; although as you have probably seen in previous posts
Teradata usage of secondary indexes is more "particular" than Oracle. If you have an index in Oracle for ordering then you will tend
to not have it in Teradata. But, those are general statements and need to be evaluated on your part.
| |
