 |
 |
Archives of the TeradataForum
Message Posted: Fri, 12 Aug 2005 @ 20:53:47 GMT
| Subj: | | Re: ROWID based join! |
| |
| From: | | Victor Sokovin |
| | A row ID join is the second part of a nested join. The first part, referred to as "nested join" in the explain text, builds a spool file of
rowIDs based on joining to the NUSI subtable rows. Those row IDs are input to a subsequent second step which performs the row ID join. The row
ID join joins a spool of rowIDs (and other data from upstream) to the actual base table rows. | |
Right; this sounds like an essential term then. I mean the second part of the nested join should not be left anonymous. It deserves a name.
What might help us to accept the term as standard, though, would be to have it at least spelled consistently in the explain plans and
documentation. The spelling seems to be entirely ad-lib.
Regards,
Victor
| |
