 |
 |
Archives of the TeradataForum
Message Posted: Mon, 22 Nov 2010 @ 10:03:04 GMT
| Subj: | | Re: Teradata's Indexing: PPI vs PI |
| |
| From: | | Faible Mou |
Dieter Noeth wrote:
| | TD's hashing is like an advanced CRC32 (it's only 4 bytes). | |
| | As collitions can't be avoided TD adds a second 4 byte value (which is actually a sequence value per hash value), the combination "row
hash" plus "uniqueness value" results in a 8 byte unique id for each row in a table. | |
JFTA, Teradata Database 12.0 now supports 1,048,576 hash buckets as well as 65,536 for a system. The larger number of buckets primarily
benefits systems with thousands of AMPs, but there is no disadvantage to using the larger number of buckets on smaller systems.
For systems upgraded to Teradata Database 12.0, the default number of hash buckets remains unchanged at 65,536 buckets. For new systems or
following a sysinit, the default is 1,048,676 buckets (5 bytes), while the "row hash" is still 8 bytes.
-- Teradata Database Release Summary, p27
cheers
| |
