datatype performance

Hi, I'm a webmaster of http://www.jivejewelry.com. Somehow the website seems slows. The developer told me that a datatype in the database design could be causing the problem. I don't believe it that is possible. Is this actually possible? Please help.

You need to be more specific. Could a database datatype cause a problem? Sure, if you store all integers as strings, that could slow thngs down. That would not be the most likely scenario.|||

Any suggestion on how I can find the specific cause?

|||

Find a page that is slow. Debug the page, using F10 to stp over statements until you find one or more that are really slow. Step into that until you find what is slowing things down.

Alternately, sprinkle code with Trace statements and turn on tracing to determine what is taking time.

|||Easiest way to see if it's the database: Turn on SQL ServerProfiler. Default options should be fine. Connect it to theproduction server and let it run for a while. Now look at querieswith high values in the Duration column. Finally, go fix them.
Note, running Profiler against a production databasecan causeperformance degredation if your server is especially hammered. Ifyou're not maxing out the server though, this probably will not be anissue.
|||A good example of Datatype for an online store is MONEY and NUMERIC, money may give you rounding problems, numeric will not but it is bigger than money. Hope this helps.|||

Ok. I found the query that got high duration. Reading the execution plan for the query, it does not makes sense. I don't know what to do. Or, how to fix it.

|||Run it in the Query Analyzer and click on execution plan which will show what you need to cut out of the query. You could also post it so someone can run it and fix it for you. Hope this helps,sql

Datatype for Primary Key

I'd like to use uniqueidentifier in my database as Primary Key
Is this a good idea?
How does it effect performance?
Plz help
ApogeeDo you mean a GUID? If so I would think this is quite a long field for a
primary key to be based on.
If you do use it, and you are generating a random one every time, I would
make sure the index isn't clustered, because you won't be inserting to the
bottom of the table.
"Apogee" <developer@.bitefish.net> wrote in message
news:exmBzrbUDHA.1680@.tk2msftngp13.phx.gbl...
> I'd like to use uniqueidentifier in my database as Primary Key
> Is this a good idea?
> How does it effect performance?
> Plz help
> Apogee
>|||What disadvantages would this have? (not clustered)
Apogee
"Ryan Breakspear" <r.breakspear@.removespamfdsltd.co.uk> wrote in message
news:eeHKXxbUDHA.1692@.TK2MSFTNGP11.phx.gbl...
> Do you mean a GUID? If so I would think this is quite a long field for a
> primary key to be based on.
> If you do use it, and you are generating a random one every time, I would
> make sure the index isn't clustered, because you won't be inserting to the
> bottom of the table.
> "Apogee" <developer@.bitefish.net> wrote in message
> news:exmBzrbUDHA.1680@.tk2msftngp13.phx.gbl...
> > I'd like to use uniqueidentifier in my database as Primary Key
> >
> > Is this a good idea?
> > How does it effect performance?
> >
> > Plz help
> >
> > Apogee
> >
> >
>|||I don't know a huge amount about this subject but I believe using a
clustered index, the data is actually stored in the order of the index. A
clustered index is therefore the fastest type. A non clustered is a normal
index, which contains the information you have included in your index, and
also a link to where the actual record is stored.
If you insert lots of records in the middle of a clustered index, the server
will have to do a certain amount or re-jigging of the data to keep it in the
actual order of the primary key.
I would suggest using a clustered index if you are using an auto increment
primary key, and a non clustered index if you are inserting random values.
If the tables are small, or without much activity, this is largely
irrelevant though.
Ryan
"Apogee" <developer@.bitefish.net> wrote in message
news:Oi88I%23bUDHA.2200@.TK2MSFTNGP11.phx.gbl...
> What disadvantages would this have? (not clustered)
> Apogee
>
> "Ryan Breakspear" <r.breakspear@.removespamfdsltd.co.uk> wrote in message
> news:eeHKXxbUDHA.1692@.TK2MSFTNGP11.phx.gbl...
> > Do you mean a GUID? If so I would think this is quite a long field for
a
> > primary key to be based on.
> >
> > If you do use it, and you are generating a random one every time, I
would
> > make sure the index isn't clustered, because you won't be inserting to
the
> > bottom of the table.
> >
> > "Apogee" <developer@.bitefish.net> wrote in message
> > news:exmBzrbUDHA.1680@.tk2msftngp13.phx.gbl...
> > > I'd like to use uniqueidentifier in my database as Primary Key
> > >
> > > Is this a good idea?
> > > How does it effect performance?
> > >
> > > Plz help
> > >
> > > Apogee
> > >
> > >
> >
> >
>|||I don't like using a GUID for an artificial primary key. It takes up 4 times
as much space as an int (16 bytes vs 4 bytes), and as the primary key is
referenced in other tables and indexes, this can add up to a large amount of
unnecessary space in your database, negatively impacting performance. You
can store more than 2 billion rows in a table when you have a IDENTITY
column starting at 1 with an INT datatype, and that is enough for most
applications. If it isn't you can always use a BIGINT (8 bytes) datatype.
Using GUIDs also makes debugging more difficult than using identity, because
humans are better at remembering 1-2-3 than at remembering 32 character
hexadecimal strings. And with IDENTITY the inserts are generated in order,
which can help debugging as well.
--
Jacco Schalkwijk MCDBA, MCSD, MCSE
Database Administrator
Eurostop Ltd.
"Apogee" <developer@.bitefish.net> wrote in message
news:exmBzrbUDHA.1680@.tk2msftngp13.phx.gbl...
> I'd like to use uniqueidentifier in my database as Primary Key
> Is this a good idea?
> How does it effect performance?
> Plz help
> Apogee
>

DataSet/DataReader Performance

Hi,

I've run some tests on fetching data from a SQL Server 2005 database using a DataSet and a DataReader.

In my database, I'm calling a stored procedure that fetches one record from a table. The Sql Server Profiler says the call takes between 20 - 50 ms (within SqlServer). The C# application and Sql Server are on the same machine.

DataSet - performance: average 575ms (for steps 3 & 4) :

1.) Create Connection

2.) Create Command

3.) Create Adapter

4.) Fill DataSet using Adapter

DataReader - performance: average 200ms (for steps 3 & 4):

1.) Create Connection

2.) Create Command

3.) Execute Reader

4.) Read with Reader

Does this performance jibe with what others have seen?

Can the performance be improved either on the DataSet or the DataReader?

Any feelings on how performance will be affected with larger record sets?

This is by design. A datareader is a readonly forward cursor which is optimized for reading data. It should be prefered whenever reading data from a datasource.

HTH, Jens Suessmeyer.

http://www.sqlserver2005.de

|||

Hi Jens,

Thanks for the reply. Are there variants that can be set to opimize the read performance of the datareader (meaning to decrease the 200ms time to respond)?

Thanks,

Andy

|||

Hi,

the only thing that would be really affecting the performance would be to decrease the retrieved data with choosing some parameters / filters in the condition.

HTH, Jens Suessmeyer-

http://www.sqlserver2005.de

DataSet Performance

Hi!
I've a DataTable with 100 rows and 100 columns. Then I'm updating each
column in each row. When I call the Update method on the DataAdapter, this
generates 10000 UPDATE statements.
Are there any solutions how I can create a better solution, which reduces
the UPDATE statements? Or how can I handle such big updates with SQL Server?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrennerUse a stored procedure to do the update. This sounds very much like you
are using a table as an array though, which isn't generally a good way
to model data in SQL.
David Portas
SQL Server MVP
--|||I've already tried it with stored procedures and the performance isn't
better.
The problem on the data model is that the 100 rows and 100 columns are
representing a fincance plan. So each column must save additional
information (like formats, formula, ...).
So I have a table for each row (called "Position") and this table references
another table which stores the columns (called "PosVal") of the row. With
this data model I've the possibility that the table "PosVal" can reference
other tables which contains the format, formulas...
Or this there any other way to model this?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner
"David Portas" <REMOVE_BEFORE_REPLYING_dportas@.acm.org> wrote in message
news:1109333983.806317.283120@.g14g2000cwa.googlegroups.com...
> Use a stored procedure to do the update. This sounds very much like you
> are using a table as an array though, which isn't generally a good way
> to model data in SQL.
> --
> David Portas
> SQL Server MVP
> --
>|||It seems like you are trying to model an abstraction ("rows", "columns"
and "formulae" from a hypothetical spreadsheet) instead of modelling
the actual data. Isn't your metadata static enough to create a proper
relational representation of it? If not then I suggest you need a
middle tier to present this data. The back end may be largely
irrelevent - I'm not sure just what benefit you are hoping to get from
using SQL Server as the data store for this.
If you do have some real data to model, then A) Normalize your tables,
B) post a CREATE TABLE statement and your stored proc. Since your proc
can update an entire row at a time I would have expected 100 updates to
outperform 10,000 but that largely depends on how you are doing the
updates and what your data looks like.
David Portas
SQL Server MVP
--|||Klaus Aschenbrenner wrote:
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter,
> this generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which
> reduces the UPDATE statements? Or how can I handle such big updates
> with SQL Server?
>
Are you using the CommandBuilder to generate the code? It probably makes
more sense to write the code yourself.
You stated in a subsequent reply that you created a stored procedure to do
the update but that it did not improve performance. Could you elaborate on
what the procedure did? I'm assuming you created a procedure that accepted
parameters for each of the 100 columns and did the update for an entire row
at a time, requiring 100 calls to the procedure instead of 10000 calls to a
procedure that did 1 column at a time...
Is that correct?
Bob Barrows
--
Microsoft MVP -- ASP/ASP.NET
Please reply to the newsgroup. The email account listed in my From
header is my spam trap, so I don't check it very often. You will get a
quicker response by posting to the newsgroup.|||Klaus,
You can use SQLXML that comes with MDAC to reduce the number of hits going
to your database. You can then send an updategram. With 2.0, you have a
graceful upgrade to the Managed SQLXML driver, so it's not a deadend.
This is obviously SQL Server specific, but Oracle has other such equivalent
solutions.
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
news:#hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>|||First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
by using the tool correctly and using the native enhancements to boost
performance. That is not to say that an RDBMS can not be fast, on the
contrary; however, that is not its chief purpose.
If performance is your ONLY concern, use a flat file or an XML file.
You use an RDBMS to MODEL THE DATA, so that others can query it in a myriad
of ways and garauntee that there results are accurate. Therefore, you have
to use the RELATIONAL rules to model your data before you build the physical
database and constrain it in order to provide DATA INTEGRITY. It is this
integrity that you build on an RDBMS system. The system is optimized for
performance, but only after providing the foundation, a relational database
properly constrained.
Sincerely,
Anthony Thomas
"Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
Hi!
I've a DataTable with 100 rows and 100 columns. Then I'm updating each
column in each row. When I call the Update method on the DataAdapter, this
generates 10000 UPDATE statements.
Are there any solutions how I can create a better solution, which reduces
the UPDATE statements? Or how can I handle such big updates with SQL Server?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner|||You must be a consultant. :)
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OIk#AyDHFHA.3352@.TK2MSFTNGP10.phx.gbl...
> First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
> by using the tool correctly and using the native enhancements to boost
> performance. That is not to say that an RDBMS can not be fast, on the
> contrary; however, that is not its chief purpose.
> If performance is your ONLY concern, use a flat file or an XML file.
> You use an RDBMS to MODEL THE DATA, so that others can query it in a
myriad
> of ways and garauntee that there results are accurate. Therefore, you
have
> to use the RELATIONAL rules to model your data before you build the
physical
> database and constrain it in order to provide DATA INTEGRITY. It is this
> integrity that you build on an RDBMS system. The system is optimized for
> performance, but only after providing the foundation, a relational
database
> properly constrained.
> Sincerely,
>
> Anthony Thomas
>
>
> --
> "Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
> news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>|||Klaus,
The rows are updated depending on the rowstate.
It can be that when you start that the rowstate are set to "added" while you
don't want to update them all. (The dataadapter.fill set them automaticly
to unchanged when you have not set the property for that to false, however
when you load them by hand, by instance using a datareader they are all set
to "added".).
You can by instance in the case of that filling with the datareader set the
rowstate of all rows to unchanged by ds.acceptchanges
Maybe this helps?
Cor|||If only consultants would be so highly critical.
Sincerely,
Anthony Thomas
"Sahil Malik" <contactmethrumyblog@.nospam.com> wrote in message
news:%23Phk7IcHFHA.2984@.TK2MSFTNGP15.phx.gbl...
You must be a consultant. :)
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OIk#AyDHFHA.3352@.TK2MSFTNGP10.phx.gbl...
> First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
> by using the tool correctly and using the native enhancements to boost
> performance. That is not to say that an RDBMS can not be fast, on the
> contrary; however, that is not its chief purpose.
> If performance is your ONLY concern, use a flat file or an XML file.
> You use an RDBMS to MODEL THE DATA, so that others can query it in a
myriad
> of ways and garauntee that there results are accurate. Therefore, you
have
> to use the RELATIONAL rules to model your data before you build the
physical
> database and constrain it in order to provide DATA INTEGRITY. It is this
> integrity that you build on an RDBMS system. The system is optimized for
> performance, but only after providing the foundation, a relational
database
> properly constrained.
> Sincerely,
>
> Anthony Thomas
>
>
> --
> "Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
> news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>sql

DataSet Performance

Hi!
I've a DataTable with 100 rows and 100 columns. Then I'm updating each
column in each row. When I call the Update method on the DataAdapter, this
generates 10000 UPDATE statements.
Are there any solutions how I can create a better solution, which reduces
the UPDATE statements? Or how can I handle such big updates with SQL Server?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at, www.anecon.com
http://weblogs.asp.net/klaus.aschenbrennerUse a stored procedure to do the update. This sounds very much like you
are using a table as an array though, which isn't generally a good way
to model data in SQL.
--
David Portas
SQL Server MVP
--|||I've already tried it with stored procedures and the performance isn't
better.
The problem on the data model is that the 100 rows and 100 columns are
representing a fincance plan. So each column must save additional
information (like formats, formula, ...).
So I have a table for each row (called "Position") and this table references
another table which stores the columns (called "PosVal") of the row. With
this data model I've the possibility that the table "PosVal" can reference
other tables which contains the format, formulas...
Or this there any other way to model this?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at, www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner
"David Portas" <REMOVE_BEFORE_REPLYING_dportas@.acm.org> wrote in message
news:1109333983.806317.283120@.g14g2000cwa.googlegroups.com...
> Use a stored procedure to do the update. This sounds very much like you
> are using a table as an array though, which isn't generally a good way
> to model data in SQL.
> --
> David Portas
> SQL Server MVP
> --
>|||It seems like you are trying to model an abstraction ("rows", "columns"
and "formulae" from a hypothetical spreadsheet) instead of modelling
the actual data. Isn't your metadata static enough to create a proper
relational representation of it? If not then I suggest you need a
middle tier to present this data. The back end may be largely
irrelevent - I'm not sure just what benefit you are hoping to get from
using SQL Server as the data store for this.
If you do have some real data to model, then A) Normalize your tables,
B) post a CREATE TABLE statement and your stored proc. Since your proc
can update an entire row at a time I would have expected 100 updates to
outperform 10,000 but that largely depends on how you are doing the
updates and what your data looks like.
--
David Portas
SQL Server MVP
--|||Klaus Aschenbrenner wrote:
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter,
> this generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which
> reduces the UPDATE statements? Or how can I handle such big updates
> with SQL Server?
>
Are you using the CommandBuilder to generate the code? It probably makes
more sense to write the code yourself.
You stated in a subsequent reply that you created a stored procedure to do
the update but that it did not improve performance. Could you elaborate on
what the procedure did? I'm assuming you created a procedure that accepted
parameters for each of the 100 columns and did the update for an entire row
at a time, requiring 100 calls to the procedure instead of 10000 calls to a
procedure that did 1 column at a time...
Is that correct?
Bob Barrows
--
Microsoft MVP -- ASP/ASP.NET
Please reply to the newsgroup. The email account listed in my From
header is my spam trap, so I don't check it very often. You will get a
quicker response by posting to the newsgroup.|||Klaus,
You can use SQLXML that comes with MDAC to reduce the number of hits going
to your database. You can then send an updategram. With 2.0, you have a
graceful upgrade to the Managed SQLXML driver, so it's not a deadend.
This is obviously SQL Server specific, but Oracle has other such equivalent
solutions.
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
news:#hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at, www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>|||First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
by using the tool correctly and using the native enhancements to boost
performance. That is not to say that an RDBMS can not be fast, on the
contrary; however, that is not its chief purpose.
If performance is your ONLY concern, use a flat file or an XML file.
You use an RDBMS to MODEL THE DATA, so that others can query it in a myriad
of ways and garauntee that there results are accurate. Therefore, you have
to use the RELATIONAL rules to model your data before you build the physical
database and constrain it in order to provide DATA INTEGRITY. It is this
integrity that you build on an RDBMS system. The system is optimized for
performance, but only after providing the foundation, a relational database
properly constrained.
Sincerely,
Anthony Thomas
"Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
Hi!
I've a DataTable with 100 rows and 100 columns. Then I'm updating each
column in each row. When I call the Update method on the DataAdapter, this
generates 10000 UPDATE statements.
Are there any solutions how I can create a better solution, which reduces
the UPDATE statements? Or how can I handle such big updates with SQL Server?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at, www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner|||You must be a consultant. :)
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OIk#AyDHFHA.3352@.TK2MSFTNGP10.phx.gbl...
> First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
> by using the tool correctly and using the native enhancements to boost
> performance. That is not to say that an RDBMS can not be fast, on the
> contrary; however, that is not its chief purpose.
> If performance is your ONLY concern, use a flat file or an XML file.
> You use an RDBMS to MODEL THE DATA, so that others can query it in a
myriad
> of ways and garauntee that there results are accurate. Therefore, you
have
> to use the RELATIONAL rules to model your data before you build the
physical
> database and constrain it in order to provide DATA INTEGRITY. It is this
> integrity that you build on an RDBMS system. The system is optimized for
> performance, but only after providing the foundation, a relational
database
> properly constrained.
> Sincerely,
>
> Anthony Thomas
>
>
> --
> "Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
> news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at, www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>|||Klaus,
The rows are updated depending on the rowstate.
It can be that when you start that the rowstate are set to "added" while you
don't want to update them all. (The dataadapter.fill set them automaticly
to unchanged when you have not set the property for that to false, however
when you load them by hand, by instance using a datareader they are all set
to "added".).
You can by instance in the case of that filling with the datareader set the
rowstate of all rows to unchanged by ds.acceptchanges
Maybe this helps?
Cor|||If only consultants would be so highly critical.
Sincerely,
Anthony Thomas
"Sahil Malik" <contactmethrumyblog@.nospam.com> wrote in message
news:%23Phk7IcHFHA.2984@.TK2MSFTNGP15.phx.gbl...
You must be a consultant. :)
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OIk#AyDHFHA.3352@.TK2MSFTNGP10.phx.gbl...
> First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
> by using the tool correctly and using the native enhancements to boost
> performance. That is not to say that an RDBMS can not be fast, on the
> contrary; however, that is not its chief purpose.
> If performance is your ONLY concern, use a flat file or an XML file.
> You use an RDBMS to MODEL THE DATA, so that others can query it in a
myriad
> of ways and garauntee that there results are accurate. Therefore, you
have
> to use the RELATIONAL rules to model your data before you build the
physical
> database and constrain it in order to provide DATA INTEGRITY. It is this
> integrity that you build on an RDBMS system. The system is optimized for
> performance, but only after providing the foundation, a relational
database
> properly constrained.
> Sincerely,
>
> Anthony Thomas
>
>
> --
> "Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
> news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at, www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>

DataSet Performance

Hi!
I've a DataTable with 100 rows and 100 columns. Then I'm updating each
column in each row. When I call the Update method on the DataAdapter, this
generates 10000 UPDATE statements.
Are there any solutions how I can create a better solution, which reduces
the UPDATE statements? Or how can I handle such big updates with SQL Server?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner
Use a stored procedure to do the update. This sounds very much like you
are using a table as an array though, which isn't generally a good way
to model data in SQL.
David Portas
SQL Server MVP
|||I've already tried it with stored procedures and the performance isn't
better.
The problem on the data model is that the 100 rows and 100 columns are
representing a fincance plan. So each column must save additional
information (like formats, formula, ...).
So I have a table for each row (called "Position") and this table references
another table which stores the columns (called "PosVal") of the row. With
this data model I've the possibility that the table "PosVal" can reference
other tables which contains the format, formulas...
Or this there any other way to model this?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner
"David Portas" <REMOVE_BEFORE_REPLYING_dportas@.acm.org> wrote in message
news:1109333983.806317.283120@.g14g2000cwa.googlegr oups.com...
> Use a stored procedure to do the update. This sounds very much like you
> are using a table as an array though, which isn't generally a good way
> to model data in SQL.
> --
> David Portas
> SQL Server MVP
> --
>
|||It seems like you are trying to model an abstraction ("rows", "columns"
and "formulae" from a hypothetical spreadsheet) instead of modelling
the actual data. Isn't your metadata static enough to create a proper
relational representation of it? If not then I suggest you need a
middle tier to present this data. The back end may be largely
irrelevent - I'm not sure just what benefit you are hoping to get from
using SQL Server as the data store for this.
If you do have some real data to model, then A) Normalize your tables,
B) post a CREATE TABLE statement and your stored proc. Since your proc
can update an entire row at a time I would have expected 100 updates to
outperform 10,000 but that largely depends on how you are doing the
updates and what your data looks like.
David Portas
SQL Server MVP
|||Klaus Aschenbrenner wrote:
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter,
> this generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which
> reduces the UPDATE statements? Or how can I handle such big updates
> with SQL Server?
>
Are you using the CommandBuilder to generate the code? It probably makes
more sense to write the code yourself.
You stated in a subsequent reply that you created a stored procedure to do
the update but that it did not improve performance. Could you elaborate on
what the procedure did? I'm assuming you created a procedure that accepted
parameters for each of the 100 columns and did the update for an entire row
at a time, requiring 100 calls to the procedure instead of 10000 calls to a
procedure that did 1 column at a time...
Is that correct?
Bob Barrows
Microsoft MVP -- ASP/ASP.NET
Please reply to the newsgroup. The email account listed in my From
header is my spam trap, so I don't check it very often. You will get a
quicker response by posting to the newsgroup.
|||Klaus,
You can use SQLXML that comes with MDAC to reduce the number of hits going
to your database. You can then send an updategram. With 2.0, you have a
graceful upgrade to the Managed SQLXML driver, so it's not a deadend.
This is obviously SQL Server specific, but Oracle has other such equivalent
solutions.
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
news:#hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>
|||First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
by using the tool correctly and using the native enhancements to boost
performance. That is not to say that an RDBMS can not be fast, on the
contrary; however, that is not its chief purpose.
If performance is your ONLY concern, use a flat file or an XML file.
You use an RDBMS to MODEL THE DATA, so that others can query it in a myriad
of ways and garauntee that there results are accurate. Therefore, you have
to use the RELATIONAL rules to model your data before you build the physical
database and constrain it in order to provide DATA INTEGRITY. It is this
integrity that you build on an RDBMS system. The system is optimized for
performance, but only after providing the foundation, a relational database
properly constrained.
Sincerely,
Anthony Thomas

"Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
Hi!
I've a DataTable with 100 rows and 100 columns. Then I'm updating each
column in each row. When I call the Update method on the DataAdapter, this
generates 10000 UPDATE statements.
Are there any solutions how I can create a better solution, which reduces
the UPDATE statements? Or how can I handle such big updates with SQL Server?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner
|||You must be a consultant.
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OIk#AyDHFHA.3352@.TK2MSFTNGP10.phx.gbl...
> First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
> by using the tool correctly and using the native enhancements to boost
> performance. That is not to say that an RDBMS can not be fast, on the
> contrary; however, that is not its chief purpose.
> If performance is your ONLY concern, use a flat file or an XML file.
> You use an RDBMS to MODEL THE DATA, so that others can query it in a
myriad
> of ways and garauntee that there results are accurate. Therefore, you
have
> to use the RELATIONAL rules to model your data before you build the
physical
> database and constrain it in order to provide DATA INTEGRITY. It is this
> integrity that you build on an RDBMS system. The system is optimized for
> performance, but only after providing the foundation, a relational
database
> properly constrained.
> Sincerely,
>
> Anthony Thomas
>
>
> --
> "Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
> news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>
|||Klaus,
The rows are updated depending on the rowstate.
It can be that when you start that the rowstate are set to "added" while you
don't want to update them all. (The dataadapter.fill set them automaticly
to unchanged when you have not set the property for that to false, however
when you load them by hand, by instance using a datareader they are all set
to "added".).
You can by instance in the case of that filling with the datareader set the
rowstate of all rows to unchanged by ds.acceptchanges
Maybe this helps?
Cor
|||If only consultants would be so highly critical.
Sincerely,
Anthony Thomas

"Sahil Malik" <contactmethrumyblog@.nospam.com> wrote in message
news:%23Phk7IcHFHA.2984@.TK2MSFTNGP15.phx.gbl...
You must be a consultant.
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OIk#AyDHFHA.3352@.TK2MSFTNGP10.phx.gbl...
> First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
> by using the tool correctly and using the native enhancements to boost
> performance. That is not to say that an RDBMS can not be fast, on the
> contrary; however, that is not its chief purpose.
> If performance is your ONLY concern, use a flat file or an XML file.
> You use an RDBMS to MODEL THE DATA, so that others can query it in a
myriad
> of ways and garauntee that there results are accurate. Therefore, you
have
> to use the RELATIONAL rules to model your data before you build the
physical
> database and constrain it in order to provide DATA INTEGRITY. It is this
> integrity that you build on an RDBMS system. The system is optimized for
> performance, but only after providing the foundation, a relational
database
> properly constrained.
> Sincerely,
>
> Anthony Thomas
>
>
> --
> "Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
> news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>

DataSet Performance

Hi!
I've a DataTable with 100 rows and 100 columns. Then I'm updating each
column in each row. When I call the Update method on the DataAdapter, this
generates 10000 UPDATE statements.
Are there any solutions how I can create a better solution, which reduces
the UPDATE statements? Or how can I handle such big updates with SQL Server?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrennerUse a stored procedure to do the update. This sounds very much like you
are using a table as an array though, which isn't generally a good way
to model data in SQL.
David Portas
SQL Server MVP
--|||I've already tried it with stored procedures and the performance isn't
better.
The problem on the data model is that the 100 rows and 100 columns are
representing a fincance plan. So each column must save additional
information (like formats, formula, ...).
So I have a table for each row (called "Position") and this table references
another table which stores the columns (called "PosVal") of the row. With
this data model I've the possibility that the table "PosVal" can reference
other tables which contains the format, formulas...
Or this there any other way to model this?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner
"David Portas" <REMOVE_BEFORE_REPLYING_dportas@.acm.org> wrote in message
news:1109333983.806317.283120@.g14g2000cwa.googlegroups.com...
> Use a stored procedure to do the update. This sounds very much like you
> are using a table as an array though, which isn't generally a good way
> to model data in SQL.
> --
> David Portas
> SQL Server MVP
> --
>|||It seems like you are trying to model an abstraction ("rows", "columns"
and "formulae" from a hypothetical spreadsheet) instead of modelling
the actual data. Isn't your metadata static enough to create a proper
relational representation of it? If not then I suggest you need a
middle tier to present this data. The back end may be largely
irrelevent - I'm not sure just what benefit you are hoping to get from
using SQL Server as the data store for this.
If you do have some real data to model, then A) Normalize your tables,
B) post a CREATE TABLE statement and your stored proc. Since your proc
can update an entire row at a time I would have expected 100 updates to
outperform 10,000 but that largely depends on how you are doing the
updates and what your data looks like.
David Portas
SQL Server MVP
--|||Klaus Aschenbrenner wrote:
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter,
> this generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which
> reduces the UPDATE statements? Or how can I handle such big updates
> with SQL Server?
>
Are you using the CommandBuilder to generate the code? It probably makes
more sense to write the code yourself.
You stated in a subsequent reply that you created a stored procedure to do
the update but that it did not improve performance. Could you elaborate on
what the procedure did? I'm assuming you created a procedure that accepted
parameters for each of the 100 columns and did the update for an entire row
at a time, requiring 100 calls to the procedure instead of 10000 calls to a
procedure that did 1 column at a time...
Is that correct?
Bob Barrows
--
Microsoft MVP -- ASP/ASP.NET
Please reply to the newsgroup. The email account listed in my From
header is my spam trap, so I don't check it very often. You will get a
quicker response by posting to the newsgroup.|||Klaus,
You can use SQLXML that comes with MDAC to reduce the number of hits going
to your database. You can then send an updategram. With 2.0, you have a
graceful upgrade to the Managed SQLXML driver, so it's not a deadend.
This is obviously SQL Server specific, but Oracle has other such equivalent
solutions.
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
news:#hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>|||First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
by using the tool correctly and using the native enhancements to boost
performance. That is not to say that an RDBMS can not be fast, on the
contrary; however, that is not its chief purpose.
If performance is your ONLY concern, use a flat file or an XML file.
You use an RDBMS to MODEL THE DATA, so that others can query it in a myriad
of ways and garauntee that there results are accurate. Therefore, you have
to use the RELATIONAL rules to model your data before you build the physical
database and constrain it in order to provide DATA INTEGRITY. It is this
integrity that you build on an RDBMS system. The system is optimized for
performance, but only after providing the foundation, a relational database
properly constrained.
Sincerely,
Anthony Thomas
"Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
Hi!
I've a DataTable with 100 rows and 100 columns. Then I'm updating each
column in each row. When I call the Update method on the DataAdapter, this
generates 10000 UPDATE statements.
Are there any solutions how I can create a better solution, which reduces
the UPDATE statements? Or how can I handle such big updates with SQL Server?
Thanks
Klaus Aschenbrenner
MVP Visual C#
www.csharp.at,www.anecon.com
http://weblogs.asp.net/klaus.aschenbrenner|||You must be a consultant.
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OIk#AyDHFHA.3352@.TK2MSFTNGP10.phx.gbl...
> First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
> by using the tool correctly and using the native enhancements to boost
> performance. That is not to say that an RDBMS can not be fast, on the
> contrary; however, that is not its chief purpose.
> If performance is your ONLY concern, use a flat file or an XML file.
> You use an RDBMS to MODEL THE DATA, so that others can query it in a
myriad
> of ways and garauntee that there results are accurate. Therefore, you
have
> to use the RELATIONAL rules to model your data before you build the
physical
> database and constrain it in order to provide DATA INTEGRITY. It is this
> integrity that you build on an RDBMS system. The system is optimized for
> performance, but only after providing the foundation, a relational
database
> properly constrained.
> Sincerely,
>
> Anthony Thomas
>
>
> --
> "Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
> news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>|||Klaus,
The rows are updated depending on the rowstate.
It can be that when you start that the rowstate are set to "added" while you
don't want to update them all. (The dataadapter.fill set them automaticly
to unchanged when you have not set the property for that to false, however
when you load them by hand, by instance using a datareader they are all set
to "added".).
You can by instance in the case of that filling with the datareader set the
rowstate of all rows to unchanged by ds.acceptchanges
Maybe this helps?
Cor|||If only consultants would be so highly critical.
Sincerely,
Anthony Thomas
"Sahil Malik" <contactmethrumyblog@.nospam.com> wrote in message
news:%23Phk7IcHFHA.2984@.TK2MSFTNGP15.phx.gbl...
You must be a consultant.
- Sahil Malik
http://codebetter.com/blogs/sahil.malik/
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OIk#AyDHFHA.3352@.TK2MSFTNGP10.phx.gbl...
> First and foremost, an RDBMS is NOT FOR PERFORMANCE. You gain performance
> by using the tool correctly and using the native enhancements to boost
> performance. That is not to say that an RDBMS can not be fast, on the
> contrary; however, that is not its chief purpose.
> If performance is your ONLY concern, use a flat file or an XML file.
> You use an RDBMS to MODEL THE DATA, so that others can query it in a
myriad
> of ways and garauntee that there results are accurate. Therefore, you
have
> to use the RELATIONAL rules to model your data before you build the
physical
> database and constrain it in order to provide DATA INTEGRITY. It is this
> integrity that you build on an RDBMS system. The system is optimized for
> performance, but only after providing the foundation, a relational
database
> properly constrained.
> Sincerely,
>
> Anthony Thomas
>
>
> --
> "Klaus Aschenbrenner" <Klaus.Aschenbrenner@.anecon.com> wrote in message
> news:%23hBwlNzGFHA.2976@.TK2MSFTNGP15.phx.gbl...
> Hi!
> I've a DataTable with 100 rows and 100 columns. Then I'm updating each
> column in each row. When I call the Update method on the DataAdapter, this
> generates 10000 UPDATE statements.
> Are there any solutions how I can create a better solution, which reduces
> the UPDATE statements? Or how can I handle such big updates with SQL
Server?
> Thanks
> Klaus Aschenbrenner
> MVP Visual C#
> www.csharp.at,www.anecon.com
> http://weblogs.asp.net/klaus.aschenbrenner
>

DataPump performance problem

I'm using an SSIS under SQL 2005; i transfer data with a DataFlow Component:

DataSource = OLEDB Source;

Destination = OLEDB Destination

The source is a View returning about 100000 records; if i execute it in query analyser it takes few seconds. But if i use it as DataPump Source it takes HOURS to execute it...

I've noticed that If i modify the SELECT query inserting a "TOP 10000000000", it takes few seconds..

Why? Am i missing something?

Could I ask what would be wrong with just always using Select TOP 100% in your query? If adding the TOP 100% explicitly speeds things up that much, then that is what I would do.|||

Netoblivion wrote:

I'm using an SSIS under SQL 2005; i transfer data with a DataFlow Component:

DataSource = OLEDB Source;

Destination = OLEDB Destination

The source is a View returning about 100000 records; if i execute it in query analyser it takes few seconds. But if i use it as DataPump Source it takes HOURS to execute it...

I've noticed that If i modify the SELECT query inserting a "TOP 10000000000", it takes few seconds..

Why? Am i missing something?

Read this article:

SELECT *... or select from a dropdown in an OLE DB Source component?
(http://blogs.conchango.com/jamiethomson/archive/2006/02/21/2930.aspx)

as this could well be the problem.

Let us know!!

Thanks

-Jamie

|||Thanks Jamie that really helped me to solve this problem!!!

Databases: Active tranzactions (_Total)

MS SQL Server 2000:
When I collected performance data about the active tranzactions counter of my SQL server I found that (_Total) instance is very large (>9999999999999) and is not equal to the sum of the partial instances by DBs.
I have SP2 and all security patches installed.
What's wrong?

Thanks!Which way do you collect the data?|||Originally posted by Coolberg
Which way do you collect the data?

As usual:
Programs=>Administrative tools=>Performance=>Performance Logs and Alerts=>Counter Logs=>New Log Settings=>Name=>Add=>SQLServer: Databases=>Active Transactions+Total=>Add=>Start