From: Muli Koppel <koppel@post.tau.ac.il>
Subject: Re: TKPROF question...
Date: 1997/03/03
Message-ID: <331B15AD.3CFB@post.tau.ac.il>#1/1
References: <3317A7FD.7630@echo-on.net>
Organization: NetVision LTD.
Reply-To: koppel@post.tau.ac.il
Newsgroups: comp.databases.oracle.server



Zoran wrote:
> 
>  I've used SQL trace and tkprof to analyze some sql that some custom
> purchased software is throwing at the DB. I'm at home now and don't
> have a sample of the tkprof output to show you but...
> 
> I have observed that many SQL statements have an execute count of 2
> and a row count of 2. Others have an execute count of 4 and a row
> count of 4. By looking at the select statement I can see that only 1
> row should have been retrieved with the particular select.
> 
> My question; is this telling me that the select sql query was issued
> by the caller 2 times?
> 
> I also see something along the lines of "library cache misses: 1". If
> indeed the select sql stmt was issued more than once, the first
> execute would result in a library cache miss, but the subsequent
> execute of the same select wouldn't. Am I interpreting this correctly?
> 
> TIA
> 
> Zoran.


hi,

I encountered the same behaviour, and it seems that the count=2 
while returned rows=1 is due to the subject of Explicit/Implicit
Cursors.
While with Explicit cursors (i.e. open -> fetch & close) the oracle
server knows exactly how many rows were fetched in a particular fetch
step (always 1 if not using array processing), the use of implicit
cursors (such as select ... into :var from ...) forces oracle server to
re-issue the command in order to find out wether one - and only one -
row corresponds the select statement.

bye