 |
 |
Archives of the TeradataForum
Message Posted: Wed, 22 Mar 2006 @ 20:30:17 GMT
| Subj: | | Re: Usage of TOP function |
| |
| From: | | Dieter Noeth |
Frank.C.Martinez wrote:
| | But what about the fact that each AMP will return rows as they become available as well? So rowid X on AMP N may come back before rowid Y
on AMP M the first time, but then when AMP N is busy doing other stuff, rowid Y will come in before the other one, 'cause AMP M isn't doing as
much as the other AMP? Right? | |
Hmmm, don't know, how it's actually implemented...
Maybe just wait until the system is really busy and then:
.SESSION 10
.REPEAT 1000
insert into xxx select top 5001 * from large_table;
select distinct * from xxx;
;-)
At least in the "Single AMP optimization" case it's probably always the same.
Unless it's utilizing "Sync Scan" :-(
Dieter
| |
