Problem: Performance difference between MSDE and SQL Express 2005

[HC]

Hello, all, I started out thinking my problems were elsewhere but as I
have worked through this I have isolated my problem, currently, as a
difference between MSDE and SQL Express 2005 (I'll just call it
Express for simplicity).

I have, to try to simplify things, put the exact same DB on two
systems, one running MSDE and one running Express. Both have 2 Ghz
processors (one Intel, one AMD), both have a decent amount of RAM
(Intel system has 1 GB, AMD system has 512 MB), and plenty of GB of
free disk space. MSDE is running on the Intel system, Express is
running on the AMD system. To keep things fair I use the exact same
DB's and query on both systems. The DB's were created on MSDE so I
sp_detach_db'd them from MSDE and then sp_attach_db'd them to Express
(this is how MS says to do a "side-by-side" upgrade, so it's
acceptable to do so). After fighting problems in performance
differences in different situations I have narrowed the problem down
to this:

Executing a simple select statement with join clause on the databases
yields a difference in execution time that is quite great. Using the
Express Management program I can run the query against either system
(MSDE or Express, the two systems are connected via crossover cable to
eliminate any network problems/issues). When running the query
against the MSDE system (which is over the network) I consistently get
<20 ms response times on the query. When running the query against
the Express installation (which is in shared memory) I consistently
get 700 ms or longer response times. Both times are for the Total
Execution Time.

The query is simply this: select db1.* from db1.owner.tablename as db1
inner join db2.owner.tablename as db2 on db1.pkey = db2.someid where
db1.criteria = 3

So, gimme all the columns from one table in one DB (local to the
installation), matching the records in another DB (also local to the
installation), where one field in the first db matches a field in the
second db and where, in the first db, one column value = 3.

The first table has a total record count of 630 records of which only
12 match the where clause. The second table has a total record count
of about 2,700 of which only 12 match up on the 12 out of 630.

Even though the data is the same and I've done the detach and attach,
and even done the sp_updatestats, the difference in execution time is
remarkable, in a bad way.

Checking the Execution Plan reveals that both queries have the same
steps, but, on the MSDE system the largest consumer in the process is
the Clustered Index Scan of the 630 record table (DB1 in my query
example), using 85%. The next big consumer is a Clustered Index Seek
against the other table (2,700 rows), using 15%.

The Execution Plan against the Express system reveals basically the
exact opposite: 27% going to the Clustered Index Scan of the 630
record DB1, and 72% going to the Clustered Index Seek of the 2,700
record DB2.

I'm sorry to be stupid but I have this information but I don't know
what to do with it. The best that I can tell from this is that this
is the source of my problems. My problems are that on my current
systems that my clients use the data is returned to them faster than
they can click the mouse and that the new system (that is, when they
chose (or are forced by attrition) to move to Vista and thus Express
2005) the screen pop is like 1.5 seconds. This creates poor user
experience. Worse, one process I allow the users to do goes from
taking 14-30 seconds to over 4 minutes (all on the same machine with
the same OS and version of my program, so it's not a machine or OS or
my app problem).

Anyway, I hope someone can shed some light on this now that I've pared
it down some.

Thanks in advance.

--HC

[Erland Sommarskog]

HC (hb*****@gte.net) writes:

The query is simply this: select db1.* from db1.owner.tablename as db1
inner join db2.owner.tablename as db2 on db1.pkey = db2.someid where
db1.criteria = 3

So, gimme all the columns from one table in one DB (local to the
installation), matching the records in another DB (also local to the
installation), where one field in the first db matches a field in the
second db and where, in the first db, one column value = 3.

The first table has a total record count of 630 records of which only
12 match the where clause. The second table has a total record count
of about 2,700 of which only 12 match up on the 12 out of 630.

Even though the data is the same and I've done the detach and attach,
and even done the sp_updatestats, the difference in execution time is
remarkable, in a bad way.

Checking the Execution Plan reveals that both queries have the same
steps, but, on the MSDE system the largest consumer in the process is
the Clustered Index Scan of the 630 record table (DB1 in my query
example), using 85%. The next big consumer is a Clustered Index Seek
against the other table (2,700 rows), using 15%.

The Execution Plan against the Express system reveals basically the
exact opposite: 27% going to the Clustered Index Scan of the 630
record DB1, and 72% going to the Clustered Index Seek of the 2,700
record DB2.

Is there any index on the criteria column? It does not sound like
that, since you get a clustered scan on that table.

But before you apply any index, can you run the queries in both servers
preceded by

SET STATISTICS PROFILE ON

and then post the output? Since the output is very wide, it's not good
if you put directly into the article body, please put it in an attachment.
I see that you post from Google. I don't know if they permit attachments,
but if they don't, maybe you could put the output on a web site and just
post a URL?

Worse, one process I allow the users to do goes from taking 14-30
seconds to over 4 minutes (all on the same machine with the same OS and
version of my program, so it's not a machine or OS or my app problem).

I suppose this process is against some completely different tables?
14 seconds on the table sizes you mentioned sounds absymal to me.
--
Erland Sommarskog, SQL Server MVP, es****@sommarskog.se

Books Online for SQL Server 2005 at
http://www.microsoft.com/technet/pro...ads/books.mspx
Books Online for SQL Server 2000 at
http://www.microsoft.com/sql/prodinf...ons/books.mspx

[HC]

On Feb 4, 6:12 am, Erland Sommarskog <[e-mail]:

HC ([e-mail]) writes:

The query is simply this: select db1.* from db1.owner.tablename as db1
inner join db2.owner.tablename as db2 on db1.pkey = db2.someid where
db1.criteria = 3

So, gimme all the columns from one table in one DB (local to the
installation), matching the records in another DB (also local to the
installation), where one field in the first db matches a field in the
second db and where, in the first db, one column value = 3.

The first table has a total record count of 630 records of which only
12 match the where clause. The second table has a total record count
of about 2,700 of which only 12 match up on the 12 out of 630.

Even though the data is the same and I've done the detach and attach,
and even done the sp_updatestats, the difference in execution time is
remarkable, in a bad way.

Checking the Execution Plan reveals that both queries have the same
steps, but, on the MSDE system the largest consumer in the process is
the Clustered Index Scan of the 630 record table (DB1 in my query
example), using 85%. The next big consumer is a Clustered Index Seek
against the other table (2,700 rows), using 15%.

The Execution Plan against the Express system reveals basically the
exact opposite: 27% going to the Clustered Index Scan of the 630
record DB1, and 72% going to the Clustered Index Seek of the 2,700
record DB2.

Is there any index on the criteria column? It does not sound like
that, since you get a clustered scan on that table.

But before you apply any index, can you run the queries in both servers
preceded by

SET STATISTICS PROFILE ON

and then post the output? Since the output is very wide, it's not good
if you put directly into the article body, please put it in an attachment.
I see that you post from Google. I don't know if they permit attachments,
but if they don't, maybe you could put the output on a web site and just
post a URL?

Worse, one process I allow the users to do goes from taking 14-30
seconds to over 4 minutes (all on the same machine with the same OS and
version of my program, so it's not a machine or OS or my app problem).

I suppose this process is against some completely different tables?
14 seconds on the table sizes you mentioned sounds absymal to me.

--
Erland Sommarskog, SQL Server MVP, [e-mail]

Books Online for SQL Server 2005 athttp://www.microsoft.com/technet/prodtechnol/sql/2005/downloads/books...
Books Online for SQL Server 2000 athttp://www.microsoft.com/sql/prodinfo/previousversions/books.mspx- Hide quoted text -

- Show quoted text -

Hmm, I thought I posted this...but it has not shown up in the last 30
minutes so I'm going to send it again, I'm sorry if it winds up
showing twice...
Hello, Erland. I have no user-defined indexes (indices?) in any of my
tables or databases. The only indexes that would exist in any of the
database tables would be those that are by default with primary key
constraints. The actual query I'm running is:

select tMO.* from caredata.dbo.tbl_medications as tM inner join
caredata_aa.dbo.tbl_medicationorders as tMO on tMO.medid = tM.pkey
where tMO.resid = 43

the join clause tMO.medid is matched against the primary key of the tM
table. There is no index for tMO.medid, nor is there one for
tMO.resid. Would it help to index them since the Clustered Index Seek
is being performed against the tbl_Medications table? In my program I
do a LEFT OUTER JOIN instead of an INNER JOIN, but the performance is
the same: slow. So, I just left it at an INNER JOIN to try to keep
things simple.

I ran the query as you asked, against both database engines and have
posted them online. I was having problems with my regular website so
I had to cobble one together for this purpose. It was done through a
web-interface so I just UUE encoded a ZIP file of the text file with
the results. Each db engine's results are in a different file. I
included the query (as seen above), and the statistics. You can find
them here: http://home.earthlink.net/~hboothe/ On that lame page
you'll see links on the left side for the pages dealing with each db
engine.

If you have any problems with the files lemme know and I'll see if I
can find another place to put them.

The 14 second process I refer to (that takes over 4 minutes on
Express) is using the same tables. What I am trying to do is to pare
this stuff down to the smallest piece that exhibits the unwanted
behavior so that the situation is not obfuscated unnecessarily. I
think that once the problem with the little query is fixed, the
problem with the big query will be fixed. The process that takes 14
seconds on these little tables in MSDE and over 4 minutes in Express
is nasty. I have multiple databases. One database is the parent of
the whole system, containing information that is relevant across each
of the other databases. You may notice in the query I've posted that
the tMO table is from one db (caredata_aa.dbo.tbl_MedicationOrders)
and tM is from another (caredata.dbo.tbl_Medications). My program
(system) can support many of the _aa databases (_aa, _ab, _ac, etc.).
The process that takes 14 seconds on MSDE is supposed to report to the
users those medications that are not in use by any medication order in
any of the databases (so, no match from tM.pkey to tMO.medid from any
tMO of which there may be as many as 4 (maximum number currently
used)). Each record must not match across each of multiple db's.

I'm clearly not a SQL Server expert, I'm just a VB programmer, but
since I have my own little business I have to do all the functions.
Since this worked on MSDE so well I thought I'd done a good job of
designing the queries and working with the DB. Obviously since it's
not working well on Express there is a good chance I've just done a
poor job of designing the DB or the queries. Maybe there's something
really simple I've overlooked. Now that I've narrowed the problem
down to one little query that does one simple join, I have eliminated
a lot of the initial potential problem points (VB6, ADO, connection
strings, recordset types, etc.). I'm going to go play with adding
indexes to the tables to see if that makes any difference.

Thank you for your help.

--HC

[HC]

On Feb 4, 12:47 pm, "HC" <[e-mail]:

On Feb 4, 6:12 am, Erland Sommarskog <[e-mail]:

HC ([e-mail]) writes:

The query is simply this: select db1.* from db1.owner.tablename as db1
inner join db2.owner.tablename as db2 on db1.pkey = db2.someid where
db1.criteria = 3

So, gimme all the columns from one table in one DB (local to the
installation), matching the records in another DB (also local to the
installation), where one field in the first db matches a field in the
second db and where, in the first db, one column value = 3.

The first table has a total record count of 630 records of which only
12 match the where clause. The second table has a total record count
of about 2,700 of which only 12 match up on the 12 out of 630.

Even though the data is the same and I've done the detach and attach,
and even done the sp_updatestats, the difference in execution time is
remarkable, in a bad way.

Checking the Execution Plan reveals that both queries have the same
steps, but, on the MSDE system the largest consumer in the process is
the Clustered Index Scan of the 630 record table (DB1 in my query
example), using 85%. The next big consumer is a Clustered Index Seek
against the other table (2,700 rows), using 15%.

The Execution Plan against the Express system reveals basically the
exact opposite: 27% going to the Clustered Index Scan of the 630
record DB1, and 72% going to the Clustered Index Seek of the 2,700
record DB2.

Is there any index on the criteria column? It does not sound like
that, since you get a clustered scan on that table.

But before you apply any index, can you run the queries in both servers
preceded by

SET STATISTICS PROFILE ON

and then post the output? Since the output is very wide, it's not good
if you put directly into the article body, please put it in an attachment.
I see that you post from Google. I don't know if they permit attachments,
but if they don't, maybe you could put the output on a web site and just
post a URL?

Worse, one process I allow the users to do goes from taking 14-30
seconds to over 4 minutes (all on the same machine with the same OS and
version of my program, so it's not a machine or OS or my app problem).

I suppose this process is against some completely different tables?
14 seconds on the table sizes you mentioned sounds absymal to me.

--
Erland Sommarskog, SQL Server MVP, [e-mail]

Books Online for SQL Server 2005 athttp://www.microsoft.com/technet/prodtechnol/sql/2005/downloads/books...
Books Online for SQL Server 2000 athttp://www.microsoft.com/sql/prodinfo/previousversions/books.mspx-Hide quoted text -

- Show quoted text -

Hmm, I thought I posted this...but it has not shown up in the last 30
minutes so I'm going to send it again, I'm sorry if it winds up
showing twice...

Hello, Erland. I have no user-defined indexes (indices?) in any of my
tables or databases. The only indexes that would exist in any of the
database tables would be those that are by default with primary key
constraints. The actual query I'm running is:

select tMO.* from caredata.dbo.tbl_medications as tM inner join
caredata_aa.dbo.tbl_medicationorders as tMO on tMO.medid = tM.pkey
where tMO.resid = 43

the join clause tMO.medid is matched against the primary key of the tM
table. There is no index for tMO.medid, nor is there one for
tMO.resid. Would it help to index them since the Clustered Index Seek
is being performed against the tbl_Medications table? In my program I
do a LEFT OUTER JOIN instead of an INNER JOIN, but the performance is
the same: slow. So, I just left it at an INNER JOIN to try to keep
things simple.

I ran the query as you asked, against both database engines and have
posted them online. I was having problems with my regular website so
I had to cobble one together for this purpose. It was done through a
web-interface so I just UUE encoded a ZIP file of the text file with
the results. Each db engine's results are in a different file. I
included the query (as seen above), and the statistics. You can find
them here:http://home.earthlink.net/~hboothe/ On that lame page
you'll see links on the left side for the pages dealing with each db
engine.

If you have any problems with the files lemme know and I'll see if I
can find another place to put them.

The 14 second process I refer to (that takes over 4 minutes on
Express) is using the same tables. What I am trying to do is to pare
this stuff down to the smallest piece that exhibits the unwanted
behavior so that the situation is not obfuscated unnecessarily. I
think that once the problem with the little query is fixed, the
problem with the big query will be fixed. The process that takes 14
seconds on these little tables in MSDE and over 4 minutes in Express
is nasty. I have multiple databases. One database is the parent of
the whole system, containing information that is relevant across each
of the other databases. You may notice in the query I've posted that
the tMO table is from one db (caredata_aa.dbo.tbl_MedicationOrders)
and tM is from another (caredata.dbo.tbl_Medications). My program
(system) can support many of the _aa databases (_aa, _ab, _ac, etc.).
The process that takes 14 seconds on MSDE is supposed to report to the
users those medications that are not in use by any medication order in
any of the databases (so, no match from tM.pkey to tMO.medid from any
tMO of which there may be as many as 4 (maximum number currently
used)). Each record must not match across each of multiple db's.

I'm clearly not a SQL Server expert, I'm just a VB programmer, but
since I have my own little business I have to do all the functions.
Since this worked on MSDE so well I thought I'd done a good job of
designing the queries and working with the DB. Obviously since it's
not working well on Express there is a good chance I've just done a
poor job of designing the DB or the queries. Maybe there's something
really simple I've overlooked. Now that I've narrowed the problem
down to one little query that does one simple join, I have eliminated
a lot of the initial potential problem points (VB6, ADO, connection
strings, recordset types, etc.). I'm going to go play with adding
indexes to the tables to see if that makes any difference.

Thank you for your help.

--HC- Hide quoted text -

- Show quoted text -

Okay, after messing with indexes and some other stuff I'm no better
off than I was. I did a select * into tbl_2 to make copies of the
tables inside the same server (Express). I then ran the query against
those two new tables and the first time I did it they were blazing
fast. I ran they query in Management Express with the stats displayed
and it ran in < 100ms several times. I ran the query against the
original tables (in the same DB) and they were similarly fast. I
dropped the second (new) tables and re-created them in the same way
and ran the query but it's slow (1000 ms). I've tried so many
angles I'm losing track of them.

So, with the copies of the tables (presumably with fresh indexes and
stats) not being any faster I thought I'd mess with the indexes.

I dropped the primary key constraint on the med orders table (the one
with 630 rows) and applied a clustered index on the resident id (which
is the piece I'm using in the WHERE clause) and a non-clustered index
on the med id (which I'm referencing in the ON clause. So, two new
indexes on what should be the items that are being sought. No
difference in performance. I ran them several times in case there
might be some caching that might make a difference but no change.

I don't know what else to index the med table on, the only criteria I
seek it on in this query is the Primary Key for the table which is, by
it's very nature, a clustered index.

--HC

[Erland Sommarskog]

HC (hb*****@gte.net) writes:

select tMO.* from caredata.dbo.tbl_medications as tM inner join
caredata_aa.dbo.tbl_medicationorders as tMO on tMO.medid = tM.pkey
where tMO.resid = 43

the join clause tMO.medid is matched against the primary key of the tM
table. There is no index for tMO.medid, nor is there one for
tMO.resid. Would it help to index them since the Clustered Index Seek
is being performed against the tbl_Medications table?

There should be an index in resid. If the major part of the queries
against tbl_medicationorders are against resid, then it's probably a
good idea to cluster on this column. (You would then have to drop
the primary key, and the reapply it as nonclustered.)

In my program I do a LEFT OUTER JOIN instead of an INNER JOIN, but the
performance is the same: slow. So, I just left it at an INNER JOIN to
try to keep things simple.

Well, since you only return data from tMO, this may be even better:

SELECT tMO.*
FROM caredata_aa.dbo.tbl_medicationorders tNO
WHERE tMO.resid = 43
AND EXISTS (SELECT *
FROM caredata.dbo.tbl_medications tM
WHERE tMO.medid = tM.pkey)

Not that it is likely to affect performance, but you would avoid
returning the same row twice from meditioncationorders.

I ran the query as you asked, against both database engines and have
posted them online. I was having problems with my regular website so
I had to cobble one together for this purpose. It was done through a
web-interface so I just UUE encoded a ZIP file of the text file with
the results. Each db engine's results are in a different file. I
included the query (as seen above), and the statistics. You can find
them here: http://home.earthlink.net/~hboothe/ On that lame page
you'll see links on the left side for the pages dealing with each db
engine.

I'm afraid that they appear as somewhat cryptic to me. I guess that I
could copy them into a text editor and try to decode them, but I'm
lazy. Couldn't you just upload the output as text files there.

The 14 second process I refer to (that takes over 4 minutes on
Express) is using the same tables. What I am trying to do is to pare
this stuff down to the smallest piece that exhibits the unwanted
behavior so that the situation is not obfuscated unnecessarily. I
think that once the problem with the little query is fixed, the
problem with the big query will be fixed. The process that takes 14
seconds on these little tables in MSDE and over 4 minutes in Express
is nasty. I have multiple databases. One database is the parent of
the whole system, containing information that is relevant across each
of the other databases. You may notice in the query I've posted that
the tMO table is from one db (caredata_aa.dbo.tbl_MedicationOrders)
and tM is from another (caredata.dbo.tbl_Medications). My program
(system) can support many of the _aa databases (_aa, _ab, _ac, etc.).
The process that takes 14 seconds on MSDE is supposed to report to the
users those medications that are not in use by any medication order in
any of the databases (so, no match from tM.pkey to tMO.medid from any
tMO of which there may be as many as 4 (maximum number currently
used)). Each record must not match across each of multiple db's.

Ehum, this design sounds dubious. I get the impression that if this was
all in the same database, and the same tables, this could be done in a
single query.

Or is there any particular reason you have multiple databases?

Anyway, it occurred to me, there is one thing to check for. Run this query:

select name from sys.databases where is_auto_close_on = 1

For all databses that appear do:

ALTER DATABASE db SET AUTO_CLOSE OFF


--
Erland Sommarskog, SQL Server MVP, es****@sommarskog.se

Books Online for SQL Server 2005 at
http://www.microsoft.com/technet/pro...ads/books.mspx
Books Online for SQL Server 2000 at
http://www.microsoft.com/sql/prodinf...ons/books.mspx

[HC]

On Feb 4, 5:38 pm, Erland Sommarskog <[e-mail]:

HC ([e-mail]) writes:

select tMO.* from caredata.dbo.tbl_medications as tM inner join
caredata_aa.dbo.tbl_medicationorders as tMO on tMO.medid = tM.pkey
where tMO.resid = 43

the join clause tMO.medid is matched against the primary key of the tM
table. There is no index for tMO.medid, nor is there one for
tMO.resid. Would it help to index them since the Clustered Index Seek
is being performed against the tbl_Medications table?

There should be an index in resid. If the major part of the queries
against tbl_medicationorders are against resid, then it's probably a
good idea to cluster on this column. (You would then have to drop
the primary key, and the reapply it as nonclustered.)

In my program I do a LEFT OUTER JOIN instead of an INNER JOIN, but the
performance is the same: slow. So, I just left it at an INNER JOIN to
try to keep things simple.

Well, since you only return data from tMO, this may be even better:

SELECT tMO.*
FROM caredata_aa.dbo.tbl_medicationorders tNO
WHERE tMO.resid = 43
AND EXISTS (SELECT *
FROM caredata.dbo.tbl_medications tM
WHERE tMO.medid = tM.pkey)

Not that it is likely to affect performance, but you would avoid
returning the same row twice from meditioncationorders.

I ran the query as you asked, against both database engines and have
posted them online. I was having problems with my regular website so
I had to cobble one together for this purpose. It was done through a
web-interface so I just UUE encoded a ZIP file of the text file with
the results. Each db engine's results are in a different file. I
included the query (as seen above), and the statistics. You can find
them here:http://home.earthlink.net/~hboothe/ On that lame page
you'll see links on the left side for the pages dealing with each db
engine.

I'm afraid that they appear as somewhat cryptic to me. I guess that I
could copy them into a text editor and try to decode them, but I'm
lazy. Couldn't you just upload the output as text files there.

The 14 second process I refer to (that takes over 4 minutes on
Express) is using the same tables. What I am trying to do is to pare
this stuff down to the smallest piece that exhibits the unwanted
behavior so that the situation is not obfuscated unnecessarily. I
think that once the problem with the little query is fixed, the
problem with the big query will be fixed. The process that takes 14
seconds on these little tables in MSDE and over 4 minutes in Express
is nasty. I have multiple databases. One database is the parent of
the whole system, containing information that is relevant across each
of the other databases. You may notice in the query I've posted that
the tMO table is from one db (caredata_aa.dbo.tbl_MedicationOrders)
and tM is from another (caredata.dbo.tbl_Medications). My program
(system) can support many of the _aa databases (_aa, _ab, _ac, etc.).
The process that takes 14 seconds on MSDE is supposed to report to the
users those medications that are not in use by any medication order in
any of the databases (so, no match from tM.pkey to tMO.medid from any
tMO of which there may be as many as 4 (maximum number currently
used)). Each record must not match across each of multiple db's.

Ehum, this design sounds dubious. I get the impression that if this was
all in the same database, and the same tables, this could be done in a
single query.

Or is there any particular reason you have multiple databases?

Anyway, it occurred to me, there is one thing to check for. Run this query:

select name from sys.databases where is_auto_close_on = 1

For all databses that appear do:

ALTER DATABASE db SET AUTO_CLOSE OFF

--
Erland Sommarskog, SQL Server MVP, [e-mail]

Books Online for SQL Server 2005 athttp://www.microsoft.com/technet/prodtechnol/sql/2005/downloads/books...
Books Online for SQL Server 2000 athttp://www.microsoft.com/sql/prodinfo/previousversions/books.mspx- Hide quoted text -

- Show quoted text -

Hello, Erland. First, I have put a variety of indexes into the
tbl_medicationorders (clustered and unclustered) with no real
difference in the Total Execution Time of the query. Second, I could
do the select as you suggest but I'm not getting any records from the
database other than the 12 I expect (for this particular select
statement, so I the modified statement would seem unnecessary. Third,
and this is the worst, I've found something that makes a difference in
the execution time of the queries every time.

Briefly, before I mention what has drastically shortened the execution
time of the query let me say that i thought it was flaky as hell so I,
thinking it might be a problem with my installation of SQL Express, re-
installed SQL Express from a fresh download from MS and reloaded the
DB's. I get the same flaky problems in query execution time.

Here's what I did after the reload, carefully mapped out just as I did
it. I rebooted after re-installing Express, just as a precaution. I
detached the databases from MSDE and attached them to Express
(different boxes, of course). I ran sp_updatestats on each DB. I ran
my query as I had previously posted with the inner join. I ran it
three times with the STATISTICS PROFILE ON and, in Management Studio I
had it show me the Execution Plan and the Client Statistics tabs. I
had noticed some really strange results, sometimes really fast,
sometimes really slow, even if I had not changed the database. I
found something that makes a difference in the query execution speed
but it's really bizarre. Here is what I got:

The average Total Execution Time for those three runs is 1,202.667
ms. Really slow and intolerable. Running this query from this box
against my MSDE system gives me responses in the 15-30 ms range.

I use the Object Explorer to locate a table in the caredata database.
I right-click on a table (one completely unrelated to the query) and
click Modify. I change NOTHING, just look at the window that shows me
the fields of the table. I click back on my query tab and, after
resetting client statistics, I run the query 3 times.

Now, the average Total Execution Time is down to 650.6667 ms. No
kidding. Same query, same data, same server, same table, EVERYTHING
is the same except now I have a tab open showing me the properties of
a table in the Caredata datbase.

Next, I open another unrelated table in the caredata_aa database (the
other DB involved in the query) in the same way (right-click, hit
Modify, change NOTHING). When I return to the tab with the query on
it and reset the client statistics and then run it three times I get
an average Total Execution Time of 62.000 ms.

What I'm seeing is that when I have an object (table) open from the
databases involved in the query (even if it's just a view of the
record layout in the Management Express console), the query runs
ridiculously faster, almost as fast as it runs against the MSDE
database.

I have reposted the data on that site for you. I didn't like how it
put line breaks in it which is why I UUE encoded it. You can take UUE
encoded stuff and just save it as a text file with a UUE extension and
programs like WinZip will read it. It's there now in plain text under
the SQL Express Results or here is a direct link: http://
home.earthlink.net/~hboothe/id2.html

I posted the results twice, once from when I had the other tables open
so the query was fast and once when the other tables were closed so
the query was long.

I don't understand why opening up the tables would make a difference.
If it was on a remote server that had to be sought out it might make
sense that the query would run faster if the server had already been
discovered but since it's running in shared memory that seems
unnecessary.

I ran the query you provided, as a nested select statement but the
response times were roughly the same as what I got with the join, on
average across 3 runs, 1,181.667ms Total Execution Time.

In all candor, the design may be dubious or outright stupid. The
design is based on a prior product which used separate sub-directories
to store the data for different clients. The idea here is that my
software product can keep information for one or more businesses.
That is, I may install my software at one company and they, as a
service, might host information for many other businesses on my
application. However, I may install the software at one company who
only hosts information for themselves. Each company that is hosted on
the system has it's data completely separate from the others. Any
information that would need to be shared and used across all the
different businesses on one installation go in the "parent" database
(which I call the repository) and any information specific to the
particular client goes in the "facility" database. I'm seeing some
ways I could have done it differently now, as a result of our
discussion, but when I started this project and product 2 years ago I
didn't consider doing it differently. Now the thing is installed
across several locations and I have another client about to go live
with a new system at the end of the month, presumably with Vista which
means no MSDE, so I'm in no position to try to re-do the DB structure/
layout/design.

Hmm, at the end you mention the auto close thingy. I just pulled the
db's with auto close and it's all the ones I attached. I ran the auto
close = burn and die (okay, I'm being goofy, I ran the alter database
db set auto_close off) and then ran my query and it was blazing fast.
That auto_close, from the name, would explain why having open
connections to the DB would fix the problem.

I'm going to reboot now (to make sure I've not done anything else that
might be affecting it) and then I'll check it and post back.

Thank you again for all your help and time.

--HC

[HC]

On Feb 4, 10:48 pm, "HC" <[e-mail]:

On Feb 4, 5:38 pm, Erland Sommarskog <[e-mail]:

HC ([e-mail]) writes:

select tMO.* from caredata.dbo.tbl_medications as tM inner join
caredata_aa.dbo.tbl_medicationorders as tMO on tMO.medid = tM.pkey
where tMO.resid = 43

the join clause tMO.medid is matched against the primary key of the tM
table. There is no index for tMO.medid, nor is there one for
tMO.resid. Would it help to index them since the Clustered Index Seek
is being performed against the tbl_Medications table?

There should be an index in resid. If the major part of the queries
against tbl_medicationorders are against resid, then it's probably a
good idea to cluster on this column. (You would then have to drop
the primary key, and the reapply it as nonclustered.)

In my program I do a LEFT OUTER JOIN instead of an INNER JOIN, but the
performance is the same: slow. So, I just left it at an INNER JOIN to
try to keep things simple.

Well, since you only return data from tMO, this may be even better:

SELECT tMO.*
FROM caredata_aa.dbo.tbl_medicationorders tNO
WHERE tMO.resid = 43
AND EXISTS (SELECT *
FROM caredata.dbo.tbl_medications tM
WHERE tMO.medid = tM.pkey)

Not that it is likely to affect performance, but you would avoid
returning the same row twice from meditioncationorders.

I ran the query as you asked, against both database engines and have
posted them online. I was having problems with my regular website so
I had to cobble one together for this purpose. It was done through a
web-interface so I just UUE encoded a ZIP file of the text file with
the results. Each db engine's results are in a different file. I
included the query (as seen above), and the statistics. You can find
them here:http://home.earthlink.net/~hboothe/On that lame page
you'll see links on the left side for the pages dealing with each db
engine.

I'm afraid that they appear as somewhat cryptic to me. I guess that I
could copy them into a text editor and try to decode them, but I'm
lazy. Couldn't you just upload the output as text files there.

The 14 second process I refer to (that takes over 4 minutes on
Express) is using the same tables. What I am trying to do is to pare
this stuff down to the smallest piece that exhibits the unwanted
behavior so that the situation is not obfuscated unnecessarily. I
think that once the problem with the little query is fixed, the
problem with the big query will be fixed. The process that takes 14
seconds on these little tables in MSDE and over 4 minutes in Express
is nasty. I have multiple databases. One database is the parent of
the whole system, containing information that is relevant across each
of the other databases. You may notice in the query I've posted that
the tMO table is from one db (caredata_aa.dbo.tbl_MedicationOrders)
and tM is from another (caredata.dbo.tbl_Medications). My program
(system) can support many of the _aa databases (_aa, _ab, _ac, etc.).
The process that takes 14 seconds on MSDE is supposed to report to the
users those medications that are not in use by any medication order in
any of the databases (so, no match from tM.pkey to tMO.medid from any
tMO of which there may be as many as 4 (maximum number currently
used)). Each record must not match across each of multiple db's.

Ehum, this design sounds dubious. I get the impression that if this was
all in the same database, and the same tables, this could be done in a
single query.

Or is there any particular reason you have multiple databases?

Anyway, it occurred to me, there is one thing to check for. Run this query:

select name from sys.databases where is_auto_close_on = 1

For all databses that appear do:

ALTER DATABASE db SET AUTO_CLOSE OFF

--
Erland Sommarskog, SQL Server MVP, [e-mail]

Books Online for SQL Server 2005 athttp://www.microsoft.com/technet/prodtechnol/sql/2005/downloads/books...
Books Online for SQL Server 2000 athttp://www.microsoft.com/sql/prodinfo/previousversions/books.mspx-Hide quoted text -

- Show quoted text -

Hello, Erland. First, I have put a variety of indexes into the
tbl_medicationorders (clustered and unclustered) with no real
difference in the Total Execution Time of the query. Second, I could
do the select as you suggest but I'm not getting any records from the
database other than the 12 I expect (for this particular select
statement, so I the modified statement would seem unnecessary. Third,
and this is the worst, I've found something that makes a difference in
the execution time of the queries every time.

Briefly, before I mention what has drastically shortened the execution
time of the query let me say that i thought it was flaky as hell so I,
thinking it might be a problem with my installation of SQL Express, re-
installed SQL Express from a fresh download from MS and reloaded the
DB's. I get the same flaky problems in query execution time.

Here's what I did after the reload, carefully mapped out just as I did
it. I rebooted after re-installing Express, just as a precaution. I
detached the databases from MSDE and attached them to Express
(different boxes, of course). I ran sp_updatestats on each DB. I ran
my query as I had previously posted with the inner join. I ran it
three times with the STATISTICS PROFILE ON and, in Management Studio I
had it show me the Execution Plan and the Client Statistics tabs. I
had noticed some really strange results, sometimes really fast,
sometimes really slow, even if I had not changed the database. I
found something that makes a difference in the query execution speed
but it's really bizarre. Here is what I got:

The average Total Execution Time for those three runs is 1,202.667
ms. Really slow and intolerable. Running this query from this box
against my MSDE system gives me responses in the 15-30 ms range.

I use the Object Explorer to locate a table in the caredata database.
I right-click on a table (one completely unrelated to the query) and
click Modify. I change NOTHING, just look at the window that shows me
the fields of the table. I click back on my query tab and, after
resetting client statistics, I run the query 3 times.

Now, the average Total Execution Time is down to 650.6667 ms. No
kidding. Same query, same data, same server, same table, EVERYTHING
is the same except now I have a tab open showing me the properties of
a table in the Caredata datbase.

Next, I open another unrelated table in the caredata_aa database (the
other DB involved in the query) in the same way (right-click, hit
Modify, change NOTHING). When I return to the tab with the query on
it and reset the client statistics and then run it three times I get
an average Total Execution Time of 62.000 ms.

What I'm seeing is that when I have an object (table) open from the
databases involved in the query (even if it's just a view of the
record layout in the Management Express console), the query runs
ridiculously faster, almost as fast as it runs against the MSDE
database.

I have reposted the data on that site for you. I didn't like how it
put line breaks in it which is why I UUE encoded it. You can take UUE
encoded stuff and just save it as a text file with a UUE extension and
programs like WinZip will read it. It's there now in plain text under
the SQL Express Results or here is a direct link: http://
home.earthlink.net/~hboothe/id2.html

I posted the results twice, once from when I had the other tables open
so the query was fast and once when the other tables were closed so
the query was long.

I don't understand why opening up the tables would make a difference.
If it was on a remote server that had to be sought out it might make
sense that the query would run faster if the server had already been
discovered but since it's running in shared memory that seems
unnecessary.

I ran the query you provided, as a nested select statement but the
response times were roughly the same as what I got with the join, on
average across 3 runs, 1,181.667ms Total Execution Time.

In all candor, the design may be dubious or outright stupid. The
design is based on a prior product which used separate sub-directories
to store the data for different clients. The idea here is that my
software product can keep information for one or more businesses.
That is, I may install my software at one company and they, as a
service, might host information for many other businesses on my
application. However, I may install the software at one company who
only hosts information for themselves. Each company that is hosted on
the system has it's data completely separate from the others. Any
information that would need to be shared and used across all the
different businesses on one installation go in the "parent" database
(which I call the repository) and any information specific to the
particular client goes in the "facility" database. I'm seeing some
ways I could have done it differently now, as a result of our
discussion, but when I started this project and product 2 years ago I
didn't consider doing it differently. Now the thing is installed
across several locations and I have another client about to go live
with a new system at the end of the month, presumably with Vista which
means no MSDE, so I'm in no position to try to re-do the DB structure/
layout/design.

Hmm, at the end you mention the auto close thingy. I just pulled the
db's with auto close and it's all the ones I attached. I ran the auto
close = burn and die (okay, I'm being goofy, I ran the alter database
db set auto_close off) and then ran my query and it was blazing fast.
That auto_close, from the name, would explain why having open
connections to the DB would fix the problem.

I'm going to reboot now (to make sure I've not done anything else that
might be affecting it) and then I'll check it and post back.

Thank you again for all your help and time.

--HC

Erland, I have rebooted after running the commands to set AUTO_CLOSE
to off and the queries are fast still. That seems to be the largest
problem. The program response time for the simple query I've posted
before is in the 0.1 down to 0.08 second range, plenty fast. The
nasty query has gone from 4 minutes or more to just over 1 minute.
That's still too long but I think that with some tweaking of the
indexes I can get that down. It's not as good as MSDE kicking it out
in 14-30 seconds, but it's WAY better than 4 minutes.

It turns out, as I'm sure you know, that AUTO_CLOSE defaults to on
ONLY for SQL Express (not for other versions of SQL 2005 (from my
reading of BOL)). Turning it off was a huge help to me and my
program.

I will play with indexes tomorrow and the next day to see if I can
speed the rest of process up, but for now, the biggest problem I was
facing (the incredible slowness of the routine work) is fixed.

Thank you very much for all your time and effort on this. I very much
appreciate it.

--HC

[Erland Sommarskog]

HC (hb*****@gte.net) writes:

It turns out, as I'm sure you know, that AUTO_CLOSE defaults to on
ONLY for SQL Express (not for other versions of SQL 2005 (from my
reading of BOL)). Turning it off was a huge help to me and my
program.

As far as I know, AUTO_CLOSE is also on by default for MSDE. Probably
some time in the dim and distant path, you learnt about this, turned
it off, and by now you have forgotten all about it.

Happens to me too. Just watch this thread. It took me a couple of
posts to recognize the problem.
--
Erland Sommarskog, SQL Server MVP, es****@sommarskog.se

Books Online for SQL Server 2005 at
http://www.microsoft.com/technet/pro...ads/books.mspx
Books Online for SQL Server 2000 at
http://www.microsoft.com/sql/prodinf...ons/books.mspx

[HC]

On Feb 5, 4:20 pm, Erland Sommarskog <[e-mail]:

HC ([e-mail]) writes:

It turns out, as I'm sure you know, that AUTO_CLOSE defaults to on
ONLY for SQL Express (not for other versions of SQL 2005 (from my
reading of BOL)). Turning it off was a huge help to me and my
program.

As far as I know, AUTO_CLOSE is also on by default for MSDE. Probably
some time in the dim and distant path, you learnt about this, turned
it off, and by now you have forgotten all about it.

Happens to me too. Just watch this thread. It took me a couple of
posts to recognize the problem.

--
Erland Sommarskog, SQL Server MVP, [e-mail]

Books Online for SQL Server 2005 athttp://www.microsoft.com/technet/prodtechnol/sql/2005/downloads/books...
Books Online for SQL Server 2000 athttp://www.microsoft.com/sql/prodinfo/previousversions/books.mspx

I know what you're saying but the auto_close is off. I checked my
databases and it's off and I'd remember this kind of hell. :) My
only offering is that I designed and built the databases using the SQL
Server Enterprise Manager that I have with SQL Server 2000 Enterprise
Edition so perhaps it is a function of whatever creates the database,
not in which data engine the database is created. At least with MSDE
2000.

I tried testing this by scripting an existing database and then using
OSQL to execute the script and make a new database on MSDE on my dev
machine (that I use all the time) but it still didn't mark the DB as
auto_close. But, you are right, MSDE is supposed to, by default, mark
the databases as auto_close on (I checked BOL). I'm not sure, but it
seems odd to me that the default behavior the BOL says I should see
isn't happening which makes me think I'm doing something wrong.

In any case, setting AUTO_CLOSE off on the DB's on Express (of which
all of mine were marked on) saved a ton of time on the queries. The
simple stuff (one join, one criterion) took 1,100 to 1,500 ms or so
repeatedly, now it takes less than 300 ms.

The problem I'm working on now is why my 2Ghz Centrino on a GB of RAM
running XP SP2 and MSDE 2000 on a 5,400 RPM disk kicks out my simple
join query, inside my program but with similar speed differences from
SQL Management Express, in about 3-5 hundreths of a second (0.03 -
0.05 seconds) but my Vista system running 2GB of RAM, Athlon 64 X2
dual core 3800+, on a 7,200 RPM disk and using SQL Express 2005 kicks
the same query, on the same data, with the same program, out in about
one tenth of a second (0.10). This isn't a huge problem in itself.
What makes this a problem are two things: 1) why any perf difference
to the negative on a faster machine with the new DB system (Express
vs. MSDE), and 2) the same tables used in that little query are used
in my big nasty query and the average time on my XP system with MSDE
is 17.433 seconds over 8 runs of the big query and on the Vista system
with Express the average time is 46.66 seconds.

Here's what I did, ran the same process on each of two machines, one
is the XP system I mention above, the other is the Vista system I
mention above. I ran the process 8 times on each system, each one
right after the other. I have incorporated a timer function in my
program that tracks how long it takes to run the process (grab the
data and pop it on screen). The time is capable of recording in
milliseconds. On the MSDE system (XP) I ran the process only 8 times
with the only index on either of the tables being the clustered
primary key index. On the Express system (Vista) I ran the process 24
times, 8 each of the following: no indexes other than the clustered
primary key index, 8 each of a clustered index on MedID (the primary
key from the table with 2,738 rows) and non-clustered indexes on
primary key and ResID, and 8 each of a clustered index on ResID, and
non-clustered on MedID and primary key. After each change of the
indexes I ran first the UPDATE STATISTICS tablewithchanges WITH
FULLSCAN and then executed SP_UPDATESTATS for that database. Each of
the three groups of 8 runs averaged about the same time, with the
lowest average of 45.7 and the highest of 46.8 seconds. Approximately
3 times longer than the time the XP and MSDE system took.

There are, perhaps, things I can do to make my process smaller or
faster, I suppose, but before I try changing what I'm doing I'd like
to understand what the difference is between MSDE and Express that is
causing the problem. My expectation from Express was that I would use
it with the databases I have currently and the same program and have
the same or better performance than what I had before. What I have is
a system that, while fast enough for the little query (one tenth of a
second is not bad) is unusable for my clients when that difference is
magnified to 45+ seconds for another. To summarize, I'm disappointed
that a process that was fast on prior version is slow on a newer
version and it makes me think, particularly in light of the info you
gave me on the AUTO_CLOSE, that there is yet something I don't know or
understand and that bugs me.

Anyway, I'm not sure there's a question in there, sorry.

Thank you, ad nauseum, for sticking with me and helping me with the
initial problem. That fix is huge to my ability to move to Vista and
Express.

--HC

[Erland Sommarskog]

HC (hb*****@gte.net) writes:

I tried testing this by scripting an existing database and then using
OSQL to execute the script and make a new database on MSDE on my dev
machine (that I use all the time) but it still didn't mark the DB as
auto_close. But, you are right, MSDE is supposed to, by default, mark
the databases as auto_close on (I checked BOL). I'm not sure, but it
seems odd to me that the default behavior the BOL says I should see
isn't happening which makes me think I'm doing something wrong.

You might have turned off auto-close for the model database as well.
model is a template for all new databases.

As for the rest of your troubles, I'm afraid that I know too little about
your system to say much useful.

--
Erland Sommarskog, SQL Server MVP, es****@sommarskog.se

Books Online for SQL Server 2005 at
http://www.microsoft.com/technet/pro...ads/books.mspx
Books Online for SQL Server 2000 at
http://www.microsoft.com/sql/prodinf...ons/books.mspx