Archives of the TeradataForum

Message Posted: Fri, 28 Aug 2009 @ 10:20:34 GMT


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

Subj:		Re: Easy way to get the latest updated row in a query

From:		Dieter Noeth

Victor Sokovin wrote:

Dieter, this is also the extra condition that you seem to be using without stating it. In general, we don't know how column-level maxima are distributed across the rows in partitions. It can happen that MAX(timeStampKey1) and MAX(sequenceNbrKey1) can sit on different ones and the JOIN in the OP's query will return no rows. A more careful rewrite of that query in OLAP terms should probably include two QUALIFY clauses. Agreed?

Yep.

It looks like there are several rows with the same timeStampKey1 but different sequenceNbrKey1 (or the other way).

Otherwise there might be no row with both max values and a single QUALIFY will return a false row.

But we need to see the actual query to judge that.

It seemed to be a general question, so this was the general answer :-)

Dieter


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

Archives

2016		2007
2015		2006
2014		2005
2013		2004
2012		2003
2011		2002
2010		2001
2009		2000
2008		1999

2009 Indexes

Jan		Jul
Feb		Aug
Mar		Sep
Apr		Oct
May		Nov
Jun		Dec

Last Modified: 15 Jun 2023