Archives of the TeradataForum

Message Posted: Thu, 24 Aug 2006 @ 19:12:24 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Seems reasonable, in the sense that merge inserting 10000 rows doesn't take twice as long as merging 5000. The more rows you merge, the more likely some of them occasionally fall into the same block and you save a few disk writes.

In terms of the UPDATEs, does your EXPLAIN show a single MERGE UPDATE step or a MERGE DELETE followed by MERGE (INSERT)?

Note that for the JI, "UPDATE" generally means DELETE/INSERT - not update in place.

	<Prev	Next>		<<First	<Prev	Next>	Last>>