SQL

When aspnet_regsql.exe won't connect

Ryan Van Slooten — Thu, 24 Apr 2008 20:50:31 GMT

I was building a quick test web site and I was using the aspnet_regsql tool to add membership to a SQLEXPRESS database. At first, I tried:

aspnet_regsql -A all -C "Data Source=.\SQLEXPRESS;Integrated Security=True;User Instance=True" -d "C:\code\Test\APP_DATA\aspnetdb.mdf"

For some reason, the SqlConnection insisted that it try to create the database and disregarded the full path to the database (note the path in the exception).

SQL Exception:
System.Data.SqlClient.SqlException: Directory lookup for the file "C:\Documents and Settings\user\Local Settings\Application Data\Microsoft\Microsoft SQL Server Data\SQLEXPRESS\C:\code\Asp.net\ServerControlTest\App_Data\Database.mdf" failed with the operating system error 123(The filename, directory name, or volume label syntax is incorrect.).
CREATE DATABASE failed. Some file names listed could not be created. Check related errors.
Creating the C:\code\Asp.net\ServerControlTest\App_Data\Database database...
   at System.Data.SqlClient.SqlConnection.OnError(SqlException exception, Boolean breakConnection)
   at System.Data.SqlClient.SqlInternalConnection.OnError(SqlException exception, Boolean breakConnection)
   at System.Data.SqlClient.TdsParser.ThrowExceptionAndWarning(TdsParserStateObject stateObj)
   at System.Data.SqlClient.TdsParser.Run(RunBehavior runBehavior, SqlCommand cmdHandler, SqlDataReader dataStream, BulkCopySimpleResultSet bulkCopyHandler, TdsParserStateObject stateObj)
   at System.Data.SqlClient.SqlCommand.RunExecuteNonQueryTds(String methodName, Boolean async)
   at System.Data.SqlClient.SqlCommand.InternalExecuteNonQuery(DbAsyncResult result, String methodName, Boolean sendToPipe)
   at System.Data.SqlClient.SqlCommand.ExecuteNonQuery()
   at System.Web.Management.SqlServices.ExecuteFile(String file, String server, String database, String dbFileName, SqlConnection connection, Boolean sessionState, Boolean isInstall, SessionStateType sessionStatetype)

Next I tried the simple commands (credit Scott Mitchell) that worked great:

sqlcmd -S localhost\SQLExpress -Q "EXEC sp_attach_db 'Foobar', N'pathToDBfile'"
aspnet_regsql.exe -S localhost\SQLExpress -d Foobar -E -A all

References

Unable to log into SQL Server after creating self-signed SSL certificate

Ryan Van Slooten — Thu, 24 Apr 2008 16:29:18 GMT

In my SharePoint experiments with form-based authentication (FBA), I have been installing self-signed SSL certificates since I am developing in a virtual machine without a certificate authority. Last night, I shut down my virtual machine instead of the usual Suspend operation. This morning, I started the virtual machine but SharePoint wasn't working and said: Cannot connect to the configuration database.

I then tried to open SQL Server Management Console and tried to connect, only to receive the following message:

I didn't freak out, but I did check the SQL Server Service which was running, and then the SQL Server Log files which looked normal. There was one line which I didn't recognize:

The certificate was successfully loaded for encryption.

This is actually a normal line in the log file, but it was then that I realized that the self-signed certificates I created were interfering with the login process. I searched around for Internet resources, but finally found what I was looking for in the Certificate MMC snap-in configuration. Since I had created multiple SSL certificates, I deleted the unused certificates and checked the encryption and certificate settings in the SQL Server Configuration Manager. After a quick reboot, I was back in business.

References:

Overcoming SQL GROUP BY challenges with ROW_NUMBER

Ryan Van Slooten — Mon, 14 Apr 2008 19:35:46 GMT

In SQL, aggregate functions will return the group by values or the aggregate function results, but it is difficult (or at least harder than it should be) to return the primary key or ROWID. In contrast, most programming languages will return the instance (or a pointer/reference to the instance) when searching for items.

Background

A project manager allocated me several weeks ago on emergency basis to help out with another project that was having difficulties with a SQL Historian system. I ended up developing a nice set of SQL tables, functions, and stored procedures to transfer data from a remote linked server into the Historian. It was actually rather fun to work on as it had several challenging aspects. Of course, after several late nights of development, I don't know how much fun like that I could take!

The key to the data transfer was obtaining the latest data from the remote database and inserting it into the Historian. The remote database was an Oracle database and the Historian was a SQL Server database. My first reaction was to do a simple query with a TOP specification but Oracle doesn't have a TOP function so I started to search for ways to achieve the same result. I came across the ROW_NUMBER() function and it did the trick although it required a sub-query such as:

SELECT  *
FROM    (
    SELECT
        SAMPLE_DAY
    ,   ASSET
    ,   TANK_LEVEL_MM
    ,   TANK_PRES_PSIG
    ,   TANK_VOL_M3
    ,   FEED_FLOW_M3HR
    ,   ROW_NUMBER() OVER (PARTITION BY ASSET ORDER BY SAMPLE_DAY DESC) RN
    FROM    REMOTE_DATA_VIEW_1
    ORDER BY ASSET, SAMPLE_DAY DESC
)
WHERE   RN = 1

It turns out that this query is more powerful than the TOP query as it essentially performs a GROUP BY aggregate (ASSET and order by SAMPLE_DAY descending) without some of the limitations of SQL aggregate functions.

Exploring Aggregate Queries

In a traditional SQL aggregate query, the columns returned by the query have to be in the GROUP BY clause or in an aggregate function. This means that it is difficult to return the primary key for a table because you usually are grouping by a name, location, or other non-unique field. Here is an example:

-- DDL: This table stores the tag names, source, and values. This
-- table is just for demonstrative purposes and does not represent
-- a normalized structure.
CREATE TABLE TAG_VALUE
(
    ValueID             int             IDENTITY    PRIMARY KEY
,   DateTime            datetime        NOT NULL
,   TagName             varchar(50)     NOT NULL
,   Value               numeric             NULL
,   Ignored             bit             NOT NULL    DEFAULT 0
,   Processed           bit             NOT NULL    DEFAULT 0
)
GO

-- Get the most recent entry grouped by tag name
SELECT  TagName
,       MAX(DateTime) AS DateTime
FROM    TAG_VALUE
GROUP BY TagName

If you try to put the primary key as a return column, you will get an error. Unfortunately this means you end up having an ugly subquery and potentially expensive (hopefully you have indices on the matching columns as well), such as:

SELECT  ValueID
FROM    TAG_VALUE   AS V1
    JOIN (

        SELECT  TagName
        ,       MAX(DateTime) AS DateTime
        FROM    TAG_VALUE
        GROUP BY TagName

    )               AS V2
    ON (    (V1.TagName  = V2.TagName)
        AND (V1.DateTime = V2.DateTime))

Using ROW_NUMBER

The ROW_NUMBER function has the powerful feature of specifying the PARTITION BY which provides some of the same functionality that GROUP BY would perform but with less hassle. For example, the following query analyzes the TAG_VALUE table and looks for duplicate TagName rows. It then marks the "old" values as "Ignored" and leaves the newest entry alone.

UPDATE  TAG_VALUE
SET     Ignored = 1
FROM    TAG_VALUE AS V1
    JOIN (
        SELECT  ValueID
        ,       ROW_NUMBER() OVER (PARTITION BY TagName ORDER BY DateTime DESC) AS RN
        FROM    TAG_VALUE
        WHERE   TagName IN (
            SELECT DISTINCT TagName
            FROM        TAG_VALUE
            GROUP BY    TagName
            HAVING      COUNT(TagName) > 1
            )
    ) AS V2
    ON (V1.ValueID = V2.ValueID)
WHERE   RN > 1

Conclusion

So, in conclusion, ROW_NUMBER provides an easier method of returning the actual row so you can perform updates, deletes, or return the primary key.

Sample SQL Code

/* ====
NOTE: The table structure resembles a database design that I have seen used.
I did not design the table structure and it is intentionally flat. Originally
this example used a remote query to allow SQL Server to query ORACLE, but for
the sample it is all in SQL Server. With very minor changes it works in Oracle
as well.
==== */

-- DDL: This table simulates a remote linked server
CREATE TABLE REMOTE_DATA_VIEW_1
(
    SAMPLE_DAY          datetime
,   ASSET               varchar(50)
,   TANK_LEVEL_MM       numeric
,   TANK_PRES_PSIG      numeric
,   TANK_VOL_M3         numeric
,   FEED_FLOW_M3HR      numeric
)
GO

-- Sample data
INSERT REMOTE_DATA_VIEW_1 VALUES ('2008-02-04', 'ASSET_01', 1200.0, 18.5, 100.0, 150.0)
INSERT REMOTE_DATA_VIEW_1 VALUES ('2008-02-01', 'ASSET_01', 1100.0, 17.5,  90.0, 145.0)
INSERT REMOTE_DATA_VIEW_1 VALUES ('2008-01-25', 'ASSET_01', 1000.0, 16.5,  80.0, 155.0)
INSERT REMOTE_DATA_VIEW_1 VALUES ('2008-02-03', 'ASSET_02',  800.0, 19.5, 110.0, 160.0)
INSERT REMOTE_DATA_VIEW_1 VALUES ('2008-02-02', 'ASSET_02',  750.0, 17.0,  95.0, 165.0)
INSERT REMOTE_DATA_VIEW_1 VALUES ('2008-02-04', 'ASSET_03', 1500.0, 18.0, 100.0, 170.0)
INSERT REMOTE_DATA_VIEW_1 VALUES ('2008-01-31', 'ASSET_04', 1350.0, 19.0,  90.0, 135.0)
GO


-- Return the raw data
SELECT
    SAMPLE_DAY
,   ASSET
,   TANK_LEVEL_MM
,   TANK_PRES_PSIG
,   TANK_VOL_M3
,   FEED_FLOW_M3HR
,   ROW_NUMBER() OVER (PARTITION BY ASSET ORDER BY SAMPLE_DAY DESC) RN
FROM    REMOTE_DATA_VIEW_1
ORDER BY ASSET, SAMPLE_DAY DESC

/* ====
SAMPLE_DAY              ASSET    TANK_LEVEL_MM TANK_PRES_PSIG TANK_VOL_M3 FEED_FLOW_M3HR RN
----------------------- -------- ------------- -------------- ----------- -------------- --
2008-02-04 00:00:00.000 ASSET_01 1200          19             100         150            1
2008-02-01 00:00:00.000 ASSET_01 1100          18             90          145            2
2008-01-25 00:00:00.000 ASSET_01 1000          17             80          155            3
2008-02-03 00:00:00.000 ASSET_02 800           20             110         160            1
2008-02-02 00:00:00.000 ASSET_02 750           17             95          165            2
2008-02-04 00:00:00.000 ASSET_03 1500          18             100         170            1
2008-01-31 00:00:00.000 ASSET_04 1350          19             90          135            1
==== */


-- Use a simple embedded query
SELECT  *
FROM    (
    SELECT
        SAMPLE_DAY
    ,   ASSET
    ,   TANK_LEVEL_MM
    ,   TANK_PRES_PSIG
    ,   TANK_VOL_M3
    ,   FEED_FLOW_M3HR
    ,   ROW_NUMBER() OVER (PARTITION BY ASSET ORDER BY SAMPLE_DAY DESC) AS RN
    FROM
        REMOTE_DATA_VIEW_1
    )   DATA
WHERE   RN = 1

/* ====
SAMPLE_DAY              ASSET    TANK_LEVEL_MM TANK_PRES_PSIG TANK_VOL_M3 FEED_FLOW_M3HR RN
----------------------- -------- ------------- -------------- ----------- -------------- --
2008-02-04 00:00:00.000 ASSET_01 1200          19             100         150            1
2008-02-03 00:00:00.000 ASSET_02 800           20             110         160            1
2008-02-04 00:00:00.000 ASSET_03 1500          18             100         170            1
2008-01-31 00:00:00.000 ASSET_04 1350          19             90          135            1
==== */


-- This is an uglier version of the query above but provided for comparison.
-- Return only the most recent entry using SAMPLE_DAY and ASSET as keys. On a production
-- table, there should be a real primary key or ROWID that you would use instead.
SELECT
    DATA.SAMPLE_DAY
,   DATA.ASSET
,   DATA.TANK_LEVEL_MM
,   DATA.TANK_PRES_PSIG
,   DATA.TANK_VOL_M3
,   DATA.FEED_FLOW_M3HR
FROM    (

        SELECT
            SAMPLE_DAY
        ,   ASSET
        ,   ROW_NUMBER() OVER (PARTITION BY ASSET ORDER BY SAMPLE_DAY DESC) AS RN
        FROM
            REMOTE_DATA_VIEW_1

    )   AS  REMOTE_QUERY
    JOIN    REMOTE_DATA_VIEW_1  AS DATA
        ON (    (REMOTE_QUERY.SAMPLE_DAY  = DATA.SAMPLE_DAY)
            AND (REMOTE_QUERY.ASSET       = DATA.ASSET) )

WHERE   REMOTE_QUERY.RN = 1
ORDER BY DATA.ASSET, DATA.SAMPLE_DAY DESC

/* ====
SAMPLE_DAY              ASSET    TANK_LEVEL_MM TANK_PRES_PSIG TANK_VOL_M3 FEED_FLOW_M3HR
----------------------- -------- ------------- -------------- ----------- --------------
2008-02-04 00:00:00.000 ASSET_01 1200          19             100         150
2008-02-03 00:00:00.000 ASSET_02 800           20             110         160
2008-02-04 00:00:00.000 ASSET_03 1500          18             100         170
2008-01-31 00:00:00.000 ASSET_04 1350          19             90          135
==== */
GO

Technorati tags: SQL, SQL Server, Oracle

Required Fields with Linked SQL Servers?

Ryan Van Slooten — Fri, 12 Oct 2007 22:01:13 GMT

I was setting up a linked server from my web database to my local database to ease the backup procedure. Currently I have a script which I run to copy all of the database from the web database to my local database. I thought I could simply use sp_addlinkedserver to conect the SQL Servers. My initial attempts to connect the servers was rebuffed:

-- Try 1 (failed)
EXEC sp_addlinkedserver 
     @server        = 'subtext_web'
    ,@srvproduct    = 'SQL Server'
    ,@datasrc       = 'some.sql-web-host.com'
    ,@catalog       = 'some-db'

-- Results in the following error message:
--   Msg 15426, Level 16, State 1, Procedure sp_addlinkedserver, Line 20
--   You must specify a provider name with this set of properties.
GO


-- Try 2 (failed)
EXEC sp_addlinkedserver 
     @server        = 'subtext_web'
    ,@srvproduct    = 'SQL Server'
    ,@provider      = 'SQLNCLI'
    ,@datasrc       = 'some.sql-web-host.com'
    ,@catalog       = 'some-db'

-- Results in the following error message:
--   Msg 15428, Level 16, State 1, Procedure sp_addlinkedserver, Line 37
--   You cannot specify a provider or any properties for product 'SQL Server'.
GO


-- Try 3 (succeeded)
EXEC sp_addlinkedserver 
     @server        = 'subtext_web'
    ,@srvproduct    = ''
    ,@provider      = 'SQLNCLI'
    ,@datasrc       = 'some.sql-web-host.com'
    ,@catalog       = 'some-db'

-- Results in the following message:
--   Command(s) completed successfully.
GO

I read the documentation on sp_addlinkedserver and there was a footnote on the use of @srvproduct = 'SQL Server':

This way of setting up a linked server forces the name of the linked server to be the same as the network name of the remote instance of SQL Server. Use data_source to specify the server.

Since I was successful in Try 3, it wasn't a big deal but the error messages in Try 1 and Try 2 seemed to contradict each other since the error message in Try 1 specifically required a provider.

Technorati tags: SQL Server

Percentage Normalization in SQL

Ryan Van Slooten — Mon, 17 Sep 2007 16:40:30 GMT

I had a little fun with SQL Server last week trying to normalize a gas composition function that was returning a molar percentage of the gas. During testing we found several cases where the sum of the components did not add up to 100.0%. Of course due to floating point inaccuracies it is quite possible that rounding errors could also contribute to the issue. The values we were seeing were a couple of percentage points different and so we decided to normalize the data to ensure that it added up to 100.0%. At the same time, I rounded the components (to two decimal places in molar %) since there are contractual documents specifying the precision that will be used during final calculation.

One interesting point is that if you simply round the components after the summation, the total percentage may not be 100% either. Here is an example using purely theoretical values:

Component	Actual	Rounded
C1	10.955556	10.96
C2	8.904444	8.90
C3	8.904444	8.90
iC4	8.904444	8.90
nC4	8.904444	8.90
iC5	8.904444	8.90
nC5	8.904444	8.90
C6+	8.904444	8.90
N2	8.904444	8.90
CO2	8.904444	8.90
O2	8.904444	8.90
Total	100.000000	99.96

To account for this, I rounded all of the values except the final (chosen) component and then applied the remainder to that component. So the final normalization and rounding might look like this:

Component	Actual	Normalized
C1	94.719559	95.00
C2	3.579182	3.59
C3	0.923634	0.93
iC4	0.343786	0.34
nC4	0.050411	0.05
iC5	0.050411	0.05
nC5	0.000000	0.00
C6+	0.000000	0.00
N2	0.040948	0.04
CO2	0.000000	0.00
O2	0.000000	0.00
Total	99.707930	100.00

IF (ABS(100.0 - @sumComponents) <= @threshold)
BEGIN
    SET @diff = 100.0 - @sumComponents
    SET @factor = 1.0 + (@diff / 100.0)
END
ELSE
BEGIN
    -- No correction factor
    SET @factor = 1.0
END

-- Round the components to 2 decimal places for molar percent
SELECT  @C2  = ROUND(@C2  * @factor, 2)
       ,@C3  = ROUND(@C3  * @factor, 2)
       ,@iC4 = ROUND(@iC4 * @factor, 2)
       ,@nC4 = ROUND(@nC4 * @factor, 2)
       ,@iC5 = ROUND(@iC5 * @factor, 2)
       ,@nC5 = ROUND(@nC5 * @factor, 2)
       ,@C6  = ROUND(@C6  * @factor, 2)
       ,@N2  = ROUND(@N2  * @factor, 2)
       ,@CO2 = ROUND(@CO2 * @factor, 2)
       ,@O2  = ROUND(@O2  * @factor, 2)

-- Apply error/remainder to methane
SELECT @diff = ROUND(@C2 + @C3 + @iC4 + @nC4 + @iC5 + @nC5 + @C6 + @N2 + @CO2 + @O2, 2)
SELECT @C1   = ROUND(100.0 - @diff, 2)

-- Verify total
SELECT  @sumComponents = (@C1 + @C2 + @C3 + @iC4 + @nC4 + @iC5 + @nC5 + @C6 + @N2 + @CO2 + @O2)

Technorati tags: SQL

Blog Maintenance Is A Pain

Ryan Van Slooten — Wed, 01 Aug 2007 05:34:38 GMT

I must be getting tired... Once again I almost damaged my blog trying to get a pristine local copy. This time I decided that it would be easiest to export the blog as BlogML and reload it into the local web application. However I didn't notice that when I first exported the blog, it generated an error (I think due to the "Embed Attachments" option) in the export so I regenerated the export without embedded attachments. When I went to import the BlogML, somehow I didn't notice that it was connected to my real blog and not my local blog (like I said it is getting late!) and imported a duplicate of the information I had already created. At least this time I was able to cover my tracks without incident. Here is the SQL I used to clean up the database:

-- I determined the last valid entry was ID = 34
-- so I removed entries after 34 and reseeded the
-- identity column.
DELETE    subtext_Links
WHERE    PostID IN (
    SELECT    ID
    FROM    subtext_Content
    WHERE    ID > 34
)
GO

DELETE    subtext_Feedback
WHERE    EntryID IN (
    SELECT    ID
    FROM    subtext_Content
    WHERE    ID > 34
)
GO

DELETE    subtext_Content
WHERE    ID > 34
GO

DBCC CHECKIDENT ('dbo.subtext_Content', RESEED, 34)
GO

I had to recheck the documentation on CHECKIDENT just to make sure that I should use the maximum value (34) and not the next value (35). According to the documentation:

Current identity value is set to the new_reseed_value. If no rows have been inserted to the table since it was created, the first row inserted after you run DBCC CHECKIDENT uses new_reseed_value as the identity. Otherwise, the next row inserted uses new_reseed_value + the current increment value.

When It Rains It Pours

Ryan Van Slooten — Tue, 31 Jul 2007 19:50:58 GMT

Aside from the fact Houston has been getting a significant amount of rain this year, it seems that work and personal projects are also experiencing their share of challenges. I wanted to adjust the skin of my blog the day before yesterday and I wanted to try it out offline on a local installation. I also wanted to make a full backup of the database which served multiple purposes including being able to test changes locally. I had made a copy of the data previously using the SQL Server 2005 Export Data DTS mechanism, but it did not contain the views, stored procedures, or relationships.

So I embarked on a task to make a complete local copy of my database since it is stored online through web hosting. Since my current project is keeping my quite busy, I had to start the process late at night and, in hind sight, perhaps that was not wise. Try as I might, I could not convince the Export mechanism to reorder the table transfer sequence to match the required dependencies on the tables. I tried multiple times but I could not find a way to specify the order of the data transfer. Since some of the table required identity insert, I also tried to extract the actual SQL used to perform the transfer (it was late at night and I didn't want to spend the time to do it by hand) but that was unsuccessful as well.

Unfortunately, during one of my failed attempts to export the data I accidentally deleted the very data I was trying to export. It was late at night and I was not a happy camper. To top it off, after this I couldn't get to sleep either because I was so bothered by deleting the data.

Fortunately, I was able to recover the data and all was not lost however it did result in downtime of my blog for a day before I could restore the data. It does seem that I will be writing the data transfer SQL by hand since the SQL Server Export tool is doing such a poor job.

Bilinear Interpolation in SQL Server

Ryan Van Slooten — Tue, 19 Jun 2007 17:17:05 GMT

The past several days I have been wrapped in relatively complex calculations in SQL Server. The process involves gas composition data from a time-series historian that is weighted by flow-rate and averaged. Then the composition is used for several lookup tables for volume correction based on the gas mixture and temperature. The lookup tables are similar to the following structure:

Molecular Mass of Mixture	-150°C	-154°C	-158°C	-160°C	-162°C	-166°C	-170°C
16.0	-0.000012	-0.000010	-0.000009	-0.000009	-0.000008	-0.000007	-0.000007
16.5	0.000135	0.000118	0.000106	0.000100	0.000094	0.000086	0.000078
17.0	0.000282	0.000245	0.000221	0.000209	0.000197	0.000179	0.000163

The algorithm tries to find the closest x and y values to the supplied parameters. If any of the parameters match the lookup table, then standard linear interpolation can be used or if both parameters match then no computation is necessary. The bilinear interpolation is used for the x and y parameters and then between the intermediate x and y interpolated values.

The SQL lookup table is defined as follows and I created a quick script to generate INSERT statements from the Excel spreadsheet source:

CREATE TABLE [dbo].[K1_Lookup](
     X          float           NOT NULL        -- Molecular mass of mixture (Xi * Mi)
    ,Y          float           NOT NULL        -- Temperature, degC
    ,Z          float           NOT NULL        -- Volume correction factor, m^3/kmol
)

The algorithm to find the lookup values is crude in SQL. I used MAX where the value is less than or equal to the coordinate. That is: find the maximum coordinate value that is less than or equal to the desired coordinate. Please pardon the use of GOTO; I prefer not to use it but it was used specifically for error handling and validation. Since the

    -- The lookup table is available in: @DataValues
    -- Get the X variable position
    DECLARE  @X0 FLOAT
            ,@X1 FLOAT

    SELECT @X0 = MAX(X) FROM @DataValues WHERE X <= @XValue
    SELECT @X1 = MIN(X) FROM @DataValues WHERE X >= @XValue
    IF (@X0 IS NULL) OR (@X1 IS NULL) RETURN NULL

    -- Get the Y variable position    
    DECLARE  @Y0 FLOAT
            ,@Y1 FLOAT

    SELECT @Y0 = MAX(Y) FROM @DataValues WHERE Y <= @YValue
    SELECT @Y1 = MIN(Y) FROM @DataValues WHERE Y >= @YValue
    IF (@Y0 IS NULL) OR (@Y1 IS NULL) RETURN NULL

    -- Get the Z variable position    
    DECLARE  @Z00 FLOAT
            ,@Z01 FLOAT
            ,@Z10 FLOAT
            ,@Z11 FLOAT

    SELECT @Z00 = Z FROM @DataValues WHERE Y = @Y0 AND X = @X0
    SELECT @Z01 = Z FROM @DataValues WHERE Y = @Y0 AND X = @X1
    SELECT @Z10 = Z FROM @DataValues WHERE Y = @Y1 AND X = @X0
    SELECT @Z11 = Z FROM @DataValues WHERE Y = @Y1 AND X = @X1

    IF (@X0 = @X1) AND (@Y0 = @Y1)
    BEGIN
        -- Exact match for X and Y look-up values
        SELECT @Result = @Z00
    END
    ELSE IF (@X0 = @X1)
    BEGIN
        -- Exact match for X look-up value; find Y value
        SELECT @Result = @Z10 + (@Z11-@Z10)/(@Y1-@Y0)*(@YValue-@Y0)
    END
    ELSE IF (@Y0 = @Y1)
    BEGIN
        -- Exact match for Y look-up value; find X value
        SELECT @Result = @Z00 + (@Z01-@Z00)/(@X1-@X0)*(@XValue-@X0)

    END
    ELSE
    BEGIN
        -- Full Bilinear interpolation; determine Z position
        DECLARE  @FA FLOAT
                ,@FB FLOAT

        SELECT @FA = @Z00 + (@Z01-@Z00)/(@X1-@X0)*(@XValue-@X0)
        SELECT @FB = @Z10 + (@Z11-@Z10)/(@X1-@X0)*(@XValue-@X0)
        SELECT @Result = @FA + (@FB-@FA)/(@Y1-@Y0)*(@YValue-@Y0)
    END

    RETURN @Result

This is one area where I am keenly interested to compare the implementation and performance in a SQL based function or stored procedure compared to a CLR-based solution. Hopefully I will get some time to create a CLR-based solution and to compare the two solutions soon.

Changing table owners in SQL Server

Ryan Van Slooten — Sun, 03 Jun 2007 19:34:54 GMT

I recently installed some tables via SQL Query Analyzer under a user id that was not 'dbo'. I needed to change the owner on the tables and wrote this simple query to generate the execute statements to change the owner. Run the query to generate the next query -- so you can review the list prior to running it.

SELECT  'exec sp_changeobjectowner ''' + name + ''', ''dbo'''
FROM    sysobjects
WHERE   type = 'U'
    AND uid != 1
ORDER BY name

The result looks something like:

exec sp_changeobjectowner 'asset', 'dbo'
exec sp_changeobjectowner 'project', 'dbo'
exec sp_changeobjectowner 'salesorder', 'dbo'

Altering SQL Server

Ryan Van Slooten — Wed, 04 Apr 2007 21:17:53 GMT

I was working on a project last week where I needed to update the columns and data types in tables in SQL Server. It is pretty quick to do this manually such as:

ALTER TABLE orders
    ALTER COLUMN orderid varchar(36) NOT NULL

I prefer this compared to the code SQL Server generates in the change script which typically renames the existing table, creates a new table, copies the data from the previous table to the new table, and then restores indices and keys. While that is appropriate for dramatic changes, it is overkill for something of this nature.

Since I love using tools and scripting, I decided to make a small Python program which would let me just specify which tables and columns to modify. This simplifies maintenance as I only have to manage the tables and columns rather than reading the entire SQL script looking for missing columns or keys.

While I was developing the Python script, I wanted to add enough SQL to avoid obvious errors such as trying to add duplicate primary keys or trying to alter columns that are already NOT NULL. I used MSDN and looked at some of the generated SQL to get the following templates:

IF EXISTS (SELECT * FROM dbo.sysobjects WHERE id = OBJECT_ID('orders') AND COLUMNPROPERTY(id, 'orderid', 'AllowsNull') = 1)
BEGIN
    PRINT 'Altering column: orders.orderid'
    ALTER TABLE orders
        ALTER COLUMN orderid varchar(36) NOT NULL
END
GO

IF EXISTS (SELECT * FROM dbo.sysobjects WHERE id = OBJECT_ID('orders') AND OBJECTPROPERTY(id, 'TableHasPrimaryKey') = 0)
BEGIN
    PRINT 'Adding primary key: orders.orderid'
    ALTER TABLE orders
        ADD CONSTRAINT PK_orders
            PRIMARY KEY (orderid)
END
GO

So the Python script used the string templates to write the output SQL and it looked like:

#!/usr/bin/env python

"""Generate a T-SQL script to set the appropriate NOT NULL
columns and PRIMARY KEY indices for the database.

TODO: Create indices for specific columns as well
"""

import string

# database properties

tables = {
    "orders":           { 'notnull': [('orderid', 'varchar(36)')],
                          'pk'     : ['orderid'] },

    "accounts":         { 'notnull': [('accountid', 'varchar(50)'),
                                      ('name', 'varchar(100)')],
                          'pk'     : ['accountid'] },

    # More tables and columns...
    }

# Functions

def generatenotnull(table, column, datatype):
    print alternotnull % { 'table':table, 'column':column, 'datatype':datatype }

def generatepk(table, columns):
    print alterpk % { 'table':table, 'columns':string.join(columns, '"') }

# SQL statements and formats

alternotnull = """IF EXISTS (SELECT * FROM dbo.sysobjects WHERE id = OBJECT_ID('%(table)s') AND COLUMNPROPERTY(id, '%(column)s', 'AllowsNull') = 1)
BEGIN
    PRINT 'Altering nullable column: %(table)s.%(column)s'
    ALTER TABLE %(table)s
        ALTER COLUMN %(column)s %(datatype)s NOT NULL
END
GO"""

alterpk = """IF EXISTS (SELECT * FROM dbo.sysobjects WHERE id = OBJECT_ID('%(table)s') AND OBJECTPROPERTY(id, 'TableHasPrimaryKey') = 0)
BEGIN
    PRINT 'Adding primary key: %(table)s.%(columns)s'
    ALTER TABLE %(table)s
        ADD CONSTRAINT PK_%(table)s
            PRIMARY KEY (%(columns)s)
END
GO"""


# Main entry
if __name__ == "__main__":
    for table in tables:
        print "/* ==== Processing %s ==== */" % table
        tabledict = tables[table]
        for column in tabledict['notnull']:
            generatenotnull(table, column[0], column[1])
        # TODO: Allow for compound primary keys
        generatepk(table, tabledict['pk'])
        print ""