Database access log

SQL 2K
We developed a VB.NET application (SQL 2k db) where users can log into the
system and extract reports. We are planning to create a report with user
list who accessed the application in last 30 days.
Can this be achieved by querying directly to the database '
Thanks
JohnHello John,
This will have to be done when your application authenticates. It will need
to create an audit of who logs in and when. This isn't a function of the
database but a function of the way your application is designed.
Aaron Weiker
http://aaronweiker.com/

> SQL 2K
> We developed a VB.NET application (SQL 2k db) where users can log into
> the system and extract reports. We are planning to create a report
> with user list who accessed the application in last 30 days.
> Can this be achieved by querying directly to the database '
> Thanks
> John|||Hi John,
You can get this information, if you would have done one of the following;
(a) Enabled logging in your VB app (the user and datetime)
(b) Logging in the database tables. Again this should have been done with
your application. (the user and datetime)
(c) Enabled Auditing in the database server.
Are you using one super user to connect to SQL Server for all the
application users or are you connecting to SQL Server for every user
accessing your VB application?
In the former case, you will always see one user at the database level.
Thanks
Yogish

Database access log

SQL 2K
We developed a VB.NET application (SQL 2k db) where users can log into the
system and extract reports. We are planning to create a report with user
list who accessed the application in last 30 days.
Can this be achieved by querying directly to the database '
Thanks
JohnHi John,
You can get this information, if you would have done one of the following;
(a) Enabled logging in your VB app (the user and datetime)
(b) Logging in the database tables. Again this should have been done with
your application. (the user and datetime)
(c) Enabled Auditing in the database server.
Are you using one super user to connect to SQL Server for all the
application users or are you connecting to SQL Server for every user
accessing your VB application?
In the former case, you will always see one user at the database level.
Thanks
Yogish

Database access log

SQL 2K
We developed a VB.NET application (SQL 2k db) where users can log into the
system and extract reports. We are planning to create a report with user
list who accessed the application in last 30 days.
Can this be achieved by querying directly to the database ?
Thanks
John
Hi John,
You can get this information, if you would have done one of the following;
(a) Enabled logging in your VB app (the user and datetime)
(b) Logging in the database tables. Again this should have been done with
your application. (the user and datetime)
(c) Enabled Auditing in the database server.
Are you using one super user to connect to SQL Server for all the
application users or are you connecting to SQL Server for every user
accessing your VB application?
In the former case, you will always see one user at the database level.
Thanks
Yogish

Database access log

SQL 2K
We developed a VB.NET application (SQL 2k db) where users can log into the
system and extract reports. We are planning to create a report with user
list who accessed the application in last 30 days.
Can this be achieved by querying directly to the database '
Thanks
JohnHi John,
You can get this information, if you would have done one of the following;
(a) Enabled logging in your VB app (the user and datetime)
(b) Logging in the database tables. Again this should have been done with
your application. (the user and datetime)
(c) Enabled Auditing in the database server.
Are you using one super user to connect to SQL Server for all the
application users or are you connecting to SQL Server for every user
accessing your VB application?
In the former case, you will always see one user at the database level.
Thanks
Yogish

Data Warehouse Structure and Indexing when using MOLAP

Hi there,

I am building a cube in AS2005 which is made up of one fact table (20 million rows) and 7 dimension tables. I am planning to use MOLAP storage and my question is this. Is it important to add indexes, primary keys, foreign keys to my data source to improve performance? I can appreciate that it might speed up processing time, but since all the data and aggregations would be stored in the cube, I can't see how it would have any imact on performance (querying the cube).

If there is a benefit to adding indexes, keys etc. then could somebody please tell the best strategy, i.e where to put them, clustered or nonclustered etc.

Many thanks,

David

I have the same opinion, with MOLAP indexes will only help when you process a cube. MOLAP is separated from the data source when the cube is finished/processed.

If you use the same datamart for Reporting Services you will benefit when running queries.

The strategy for index is clustered indexes for dimension primary keys and non-clustered on their foreign keys in the fact table. A non clustered composite index on all the foreigh keys in the fact table is also recommended. Clustered index in the fact table is not recommended.

You will pay for using indexes when you do inserts, deletes and updates so it is often recommended to drop indexes when you update your data mart or data warehouse, and recreate the when the update/load is finished.

Primary and foreign keys are good help if you like to receive errors when you are loading the data mart/warehouse. With them you know that all fact records have related dimension members Else these errors will appear when you process the cube.

Here is a web site I can recommend for performance issues:

http://www.sql-server-performance.com/default.asp

Regards

Thomas Ivarsson

Data Warehouse Structure and Indexing when using MOLAP

Hi there,

I am building a cube in AS2005 which is made up of one fact table (20 million rows) and 7 dimension tables. I am planning to use MOLAP storage and my question is this. Is it important to add indexes, primary keys, foreign keys to my data source to improve performance? I can appreciate that it might speed up processing time, but since all the data and aggregations would be stored in the cube, I can't see how it would have any imact on performance (querying the cube).

If there is a benefit to adding indexes, keys etc. then could somebody please tell the best strategy, i.e where to put them, clustered or nonclustered etc.

Many thanks,

David

I have the same opinion, with MOLAP indexes will only help when you process a cube. MOLAP is separated from the data source when the cube is finished/processed.

If you use the same datamart for Reporting Services you will benefit when running queries.

The strategy for index is clustered indexes for dimension primary keys and non-clustered on their foreign keys in the fact table. A non clustered composite index on all the foreigh keys in the fact table is also recommended. Clustered index in the fact table is not recommended.

You will pay for using indexes when you do inserts, deletes and updates so it is often recommended to drop indexes when you update your data mart or data warehouse, and recreate the when the update/load is finished.

Primary and foreign keys are good help if you like to receive errors when you are loading the data mart/warehouse. With them you know that all fact records have related dimension members Else these errors will appear when you process the cube.

Here is a web site I can recommend for performance issues:

http://www.sql-server-performance.com/default.asp

Regards

Thomas Ivarsson

Data warehouse data refresh/update

Hello,
Im a member of a team planning a data warehousing project. We have multiple data sources which are aggregated in a staging area. This is then denormalised and imported into the datawarehouse database.
I am looking at ideas for incremental data refresh, rather than a drop and re-import of all data. This would allow us to have historic data.
Does anyone have any tips that might be helpful for detecting changes in the source data for import? We have had some bad experiences with triggers on our source database in the past, so would rather not use these. I have considered replication and log sh
ipping, but these just give a replica of the source data and does not flag the updated/new data.
Any help would be greatly appreciated.
Thanks.
Ben.
Look in SQL Books Online for CHECKSUM_AGG to identify changes in a table. The difficulty then is in trying to identify what has changed. Another route is to try and identify fields within the production data that will identify when it was last changed. Ty
pically these tend to be datetime, timestamp or rowversion data types.
If you keep production keys in your fact table as additional attributes then this will give you another option for identifying new data.
"Ben" wrote:

> Hello,
> Im a member of a team planning a data warehousing project. We have multiple data sources which are aggregated in a staging area. This is then denormalised and imported into the datawarehouse database.
> I am looking at ideas for incremental data refresh, rather than a drop and re-import of all data. This would allow us to have historic data.
> Does anyone have any tips that might be helpful for detecting changes in the source data for import? We have had some bad experiences with triggers on our source database in the past, so would rather not use these. I have considered replication and log
shipping, but these just give a replica of the source data and does not flag the updated/new data.
> Any help would be greatly appreciated.
> Thanks.
> Ben.
|||Timestamp your source data when it gets changed. That is definately the best way.
Rgards
Jamie
"Ben" wrote:

> Hello,
> Im a member of a team planning a data warehousing project. We have multiple data sources which are aggregated in a staging area. This is then denormalised and imported into the datawarehouse database.
> I am looking at ideas for incremental data refresh, rather than a drop and re-import of all data. This would allow us to have historic data.
> Does anyone have any tips that might be helpful for detecting changes in the source data for import? We have had some bad experiences with triggers on our source database in the past, so would rather not use these. I have considered replication and log
shipping, but these just give a replica of the source data and does not flag the updated/new data.
> Any help would be greatly appreciated.
> Thanks.
> Ben.

Data warehouse data refresh/update

Hello,
Im a member of a team planning a data warehousing project. We have multiple
data sources which are aggregated in a staging area. This is then denormalis
ed and imported into the datawarehouse database.
I am looking at ideas for incremental data refresh, rather than a drop and r
e-import of all data. This would allow us to have historic data.
Does anyone have any tips that might be helpful for detecting changes in the
source data for import? We have had some bad experiences with triggers on o
ur source database in the past, so would rather not use these. I have consid
ered replication and log sh
ipping, but these just give a replica of the source data and does not flag t
he updated/new data.
Any help would be greatly appreciated.
Thanks.
Ben.Look in SQL Books Online for CHECKSUM_AGG to identify changes in a table. Th
e difficulty then is in trying to identify what has changed. Another route i
s to try and identify fields within the production data that will identify w
hen it was last changed. Ty
pically these tend to be datetime, timestamp or rowversion data types.
If you keep production keys in your fact table as additional attributes then
this will give you another option for identifying new data.
"Ben" wrote:

> Hello,
> Im a member of a team planning a data warehousing project. We have multipl
e data sources which are aggregated in a staging area. This is then denormal
ised and imported into the datawarehouse database.
> I am looking at ideas for incremental data refresh, rather than a drop and
re-import of all data. This would allow us to have historic data.
> Does anyone have any tips that might be helpful for detecting changes in the sourc
e data for import? We have had some bad experiences with triggers on our source data
base in the past, so would rather not use these. I have considered replication and l
og
shipping, but these just give a replica of the source data and does not flag the updated/new
data.
> Any help would be greatly appreciated.
> Thanks.
> Ben.|||Timestamp your source data when it gets changed. That is definately the best
way.
Rgards
Jamie
"Ben" wrote:

> Hello,
> Im a member of a team planning a data warehousing project. We have multipl
e data sources which are aggregated in a staging area. This is then denormal
ised and imported into the datawarehouse database.
> I am looking at ideas for incremental data refresh, rather than a drop and
re-import of all data. This would allow us to have historic data.
> Does anyone have any tips that might be helpful for detecting changes in the sourc
e data for import? We have had some bad experiences with triggers on our source data
base in the past, so would rather not use these. I have considered replication and l
og
shipping, but these just give a replica of the source data and does not flag the updated/new
data.
> Any help would be greatly appreciated.
> Thanks.
> Ben.