Explain Plan problem
Posted in 1999
Topics: Performance & Tuning
Anybody run into this? When I run an explain on a query the
sqexplain.out will
not show any tables that will be accessed sequentially. But when I
actually run the query and do onstat -p the sequential scans will begin
to grow. I run a look up of the tables that have sequential scans in
sysmaster:sysptprof and one of my tables is in there with a whole bunch
of sequential scans. Does explain show the actually query path for the
query or is it the best guess?
Tim Lopez wrote:
>
> Anybody run into this? When I run an explain on a query the
> sqexplain.out will
> not show any tables that will be accessed sequentially. But when I
> actually run the query and do onstat -p the sequential scans will begin
> to grow. I run a look up of the tables that have sequential scans in
> sysmaster:sysptprof and one of my tables is in there with a whole bunch
> of sequential scans. Does explain show the actually query path for the
> query or is it the best guess?
When we upgraded to 7.30.uc3, I noticed that my seqscans increased
without my users complaining; after all, 150000 seqscans on a 1.2M row
table should be noticeable. I found that there is a bug that reports
sequential scans when Informix uses a key-first index read.
7.30.uc10 has fixed this particular issue . . . however, it's raised
another one for my company.
--
John Carlson
Informix DBA
WHSmith USA
#include std_disclaimer.h /* These are my opinions, not my company's
opinion */
In article <38517CC5.D209F202@heb.com>, Tim Lopez <pgtx@heb.com> writes
>Anybody run into this? When I run an explain on a query the
>sqexplain.out will
>not show any tables that will be accessed sequentially. But when I
>actually run the query and do onstat -p the sequential scans will begin
>to grow. I run a look up of the tables that have sequential scans in
>sysmaster:sysptprof and one of my tables is in there with a whole bunch
>of sequential scans. Does explain show the actually query path for the
>query or is it the best guess?
>
Known bug on early 7.31 where Key-First Index usage is reports as
a sequential scan..
--
David Williams