Method R Discussion Group

Method R helps developers, DBAs, and decision-makers optimize Oracle-based software in every phase of the software life cycle.

Discuss Method R products, techniques, and events.

Ask new questions or post your own responses.

Row Source - Webinar Question on 3/27/07

Row source can also be seen from DB file sequential read, correct ??
Becky Goodman Send private email
Tuesday, March 27, 2007
 
 
If, by row source, you mean the row source execution statistics provided on the STAT lines, the answer is no.  What you get from the timed event line for a "db file" read event is the file/block being read.  In other words, it shows you the "source" block(s).  The row source execution statistics show you how many block accesses (LIO and PIO) were made during a particular plan operation.

See the difference:

WAIT #15: nam='db file sequential read' ela= 15850 file#=4 block#=19900 blocks=1 obj#=59163 tim=174947956779

STAT #15 id=1 cnt=0 pid=0 pos=1 obj=59162 op='TABLE ACCESS BY INDEX ROWID EMP (cr=1 pr=1 pw=0 time=16087 us)'
STAT #15 id=2 cnt=0 pid=1 pos=1 obj=59163 op='INDEX RANGE SCAN EMP_EMPNO_PK (cr=1 pr=1 pw=0 time=16070 us)'

It's really a difference in showing the actual block accesses vs. the number of overall block accesses.  In this example trace data, it somewhat matches in that you see 1 block (block 19900) is accessed and the STAT lines show cr=1 and pr=1.  If that particular block had already been in the buffer cache, then the WAIT line wouldn't have been present and the STAT line would have shown only cr=1.

So, in the end, it depends on what you mean exactly by "row source", but I think the way Mahesh discussed it, the row source was meant to describe the STAT line(s) information and is slightly different than what the WAIT line(s) emit.
Karen Morton Send private email
Thursday, March 29, 2007
 
 

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
 
Discussion Groups Main Page

Powered by FogBugz