Hi,
when optimazer looks at the query with bind variables.. in simple terms.. it
'estimates' value of the variable, puts this into query, and then based on
different combinations vs. cost , finds the best query plan. This is all
good under assumption that data is fairly evenly distributed, so are your
variable values when they are being used in your query. Problems appear when
, distribution is havily scewed, data not indexed, or wrong/missing index on
your table(s).
To clean it up, you need to do some logging in your profiler, grab sql, try
to figure out what the query plan really should be, if you can modify sql
great if not, or play with index. Don't give up ... it has to work.
Lucjan
"Daniel Roy" <da*************@hotmail.com> wrote in message
news:37************************@posting.google.com ...
Hi gurus,
I just started to look at a very slow-running SQL statement
generated by an application (Siebel). I spooled the SQL from the
application, replaced the bind variables by their values, and tuned
from the Query Analyser. But after awhile, I realized that the
statement using bind variables and the same statement using the values
instead of the bind variables often have completely different
execution plans! Is that normal? Can someone tell me how the SQL
Server treats bind variables. Don't worry about being too technical,
I'm an Oracle DBA/developer.
Thanx
Daniel
