Friday, August 11, 2006

Use The Sieve of Eratosthenes To Find All PrimeNumbers Below 1 Million

In mathematics, the Sieve of Eratosthenes is a simple, ancient algorithm for finding all prime numbers up to a specified integer. It was created by Eratosthenes, an ancient Greek mathematician. Wheel factorization is often applied on the list of integers to be checked for primality, before Sieve of Eratosthenes is used, to increase the speed.


Algorithm
Write a list of numbers from 2 to the largest number you want to test for primality. Call this List A. (This is the list of squares on the left-hand-side of the picture.)
Write the number 2, the first prime number, in another list for primes found. Call this List B. (This is the list on the right-hand-side of the picture.)
Strike off 2 and all multiples of 2 from List A.
The first remaining number in the list is a prime number. Write this number into List B.
Strike off this number and all multiples of this number from List A. The crossing-off of multiples can be started at the square of the number, as lower multiples have already been crossed out in previous steps.
Repeat steps 4 through 6 until no more numbers are left in List A.

Read more on Wikipedia


So here is the SQL Version, it runs in 7 seconds on my machine when I run it a
second time, first run is 16 seconds


SET NOCOUNT ON


DECLARE @i INT

-- Create a 10-digit table
DECLARE @D TABLE (N INT)
INSERT INTO @D (N)
SELECT 0 UNION ALL
SELECT 1 UNION ALL
SELECT 2 UNION ALL
SELECT 3 UNION ALL
SELECT 4

INSERT INTO @D (N)
SELECT N+5 FROM @D

-- build a small sieve table between 2 and 1000
DECLARE @T TABLE (N INT)
INSERT INTO @T( N )
SELECT 1+A.N+10*(B.N+10*C.N)
FROM @D A, @D B, @D C

DELETE FROM @T WHERE N = 1

SET @I = 2
WHILE @I <= SQRT(1000)
BEGIN
DELETE FROM @T WHERE N % @I = 0 AND N > @I
SET @I = @I + 1
END

-- Create large table between 1001 and 1000000
SELECT A+10*(B+10*(C+10*(D+10*(E+ 10*F)))) AS N
INTO #P
FROM
( SELECT A.N AS A, B.N AS B, C.N AS C, D.N AS D, E.N AS E, F.N AS F
FROM @D A, @D B, @D C, @D D, @D E, @D F
WHERE A.N in (1, 3, 7, 9) -- Not divisible by 2 or 5
) blah
WHERE (A+B+C+D+E+F) % 3 <> 0 -- Or 3
AND (A+3*B+2*C-D-3*E-2*F) % 7 <> 0 -- Or 7
AND (B-A+D-C+F-E) % 11 <> 0 -- Or 11
AND D|E|F <> 0 -- Don't include the first 1000 numbers,
--we already have these in the small sieve table
UNION ALL SELECT 1000000

-- sieve the big table with smaller one
SELECT @I = 2
WHILE @I IS NOT NULL
BEGIN
DELETE FROM #P WHERE N% @I = 0
SELECT @I = MIN(N) FROM @T WHERE N > @I
END

-- add primes up to 1000
INSERT INTO #P SELECT N FROM @T

-- Here are the results
--78498 rows
SELECT * FROM #P ORDER BY 1

drop table #P
go

Wednesday, August 09, 2006

Try Quest LiteSpeed And Win An iPod Nano (Hopefully)

Quest has a nice little give away. They are giving away 40 2GB Apple Nano iPods to people who are willing to try LiteSpeed for SQL Server.
I am already using LiteSpeed on the production machines but the cool thing is that Quest has a free version that you can use with the developer editions of SQL Server at no cost at all
So check it out here

Monday, August 07, 2006

SQL Server Book Sales Up 98% Compared To Last Year

Tim O’Reilly has posted his latest State of the Computer Book Market article. Here are 2 excerpts: “Looking at the Database rollup, we again see the strength of SQL Server, the decline of Oracle book sales, and that while MySQL is still a much larger category than Postgres, Postgres is showing some curious strength. This is one of the things that treemap visualizations are good for. Small, bright green categories stand out, and you can start paying closer attention. (Ruby also showed bright green while it was still a tiny category before its remarkable surge over the past year.) We also see the continuing popularity of personal databases like Access and Filemaker.



“A few high-level observations:
• C# book sales continues to gain on Java, with a 49% unit sales increase compared to Java's 10% decrease.
• Ruby continues its momentum, and is the fastest-growing programming language in terms of book sales.
• Microsoft's new release of SQL Server has continued to drive significant book sales, with that market up 86%. ASP.Net is also on a roll, with book sales up 61%.

Read the whole article here

Thursday, August 03, 2006

Store The Output Of A Stored Procedure In A Table Without Creating A Table

I saw this technique in the Inside Microsoft SQL Server 2005: T-SQL Querying T-SQL and today I saw it also here The OPENROWSET Trick: Accessing Stored Procedure Output In A SELECT Statement @ Ward Pond's SQL Server blog

I decided to expand this a little
Sometimes I need to quickly find out who is blocking what (or what is blocking who for that matter) so I run sp_who2 then look at the BlkBy column and run a DBCC INPUTBUFFER or fn_get_sql to get the sql statement

When you have a lot of connections it's a pain in the neck to look for the BlkBy where it's not . since the result is ordered by SPID
You can always do
CREATE TABLE
INSERT TABLE
EXEC Proc

But who wants to create tables all the time (not me, at least not for this stuff)
SPID is twice in the resultset of sp_who2 (who know why?) so that complicates things a little
So let's start with sp_who


SELECT * INTO #TempSpWho
FROM OPENROWSET ('SQLOLEDB','Server=(local);TRUSTED_CONNECTION=YES;','set fmtonly off exec master.dbo.sp_who')

--return only the connections to master
SELECT * FROM #TempSpWho
WHERE dbname ='master'



--Let's try sp_who2
SELECT * INTO #TempSpWho2
FROM OPENROWSET ('SQLOLEDB','Server=(local);TRUSTED_CONNECTION=YES;','set fmtonly off exec master.dbo.sp_who2')

--Oops
Server: Msg 2705, Level 16, State 3, Line 1
Column names in each table must be unique. Column name 'SPID' in table '#TempSpWho2' is specified more than once.

--No problem list the columns
SELECT SPID, Status, Login, HostName, BlkBy,DBName, Command, CPUTime, DiskIO, LastBatch, ProgramName
INTO #TempSpWho2
FROM OPENROWSET ('SQLOLEDB','Server=(local);TRUSTED_CONNECTION=YES;','set fmtonly off exec master.dbo.sp_who2')

--Get the connections to master only
SELECT * FROM #TempSpWho2
WHERE dbname ='master'

--Get the blocking SPID's
SELECT * FROM #TempSpWho2
WHERE BlkBy NOT LIKE '% .'

--Get user connections only (everything with a SPID below 51 is reserved for sql server processes)
SELECT * FROM #TempSpWho2
WHERE SPID >= 50

A couple of notes
The OPENROWSET call opens a separate connection to SQL Server, so there is some overhead associated with this approach
This technique is not supported inside a declared transaction
Blocking can occure with a poorly architected stored procedure

Wednesday, August 02, 2006

Top SQL Server Google Searches For July 2006

These are the top SQL Searches on this site for the month of July. I have left out searches that have nothing to do with SQL Server or programming. As you can see it has been a slow(er) month. And Google reports only searches that have been searched for more than once. For example if you have these 4 searches
insert triggers
insert trigger
insert triggers
insert trigger!

Then only insert triggers will show up since the other two were used only one time


Here are the results...

increase maximum row size in mssql
crystal report division by zero code
insert triggers
select sum where
select from
delete a null record


Let's talk about the last search (delete a null record ) a little bit
The code below will show you how exactly you delete a row with a NULL value. You can not use =NULL but you have to use IS NULL

CREATE TABLE #TestDeleteNull (id INT identity, SomeDate DATETIME)
INSERT #TestDeleteNull VALUES(GETDATE())
INSERT #TestDeleteNull VALUES(GETDATE())
INSERT #TestDeleteNull VALUES(NULL)


--Check what's in the table
SELECT * FROM #TestDeleteNull

--This won't work
DELETE #TestDeleteNull
WHERE SomeDate = NULL
--(0 row(s) affected)

--This is the way
DELETE #TestDeleteNull
WHERE SomeDate IS NULL
--(1 row(s) affected)

--Check again
SELECT * FROM #TestDeleteNull

Tuesday, August 01, 2006

6 Different Ways To Get The Current Identity Value From A Table

This article will show you how to get the current identity value from a table and also some things that might act a little different than you would expect

Let's first create our two simple tables
CREATE TABLE TestOne (id INT identity,SomeDate DATETIME)
CREATE TABLE TestTwo (id INT identity,TestOneID INT,SomeDate DATETIME)

--Let's insert 4 rows into the table
INSERT TestOne VALUES(GETDATE())
INSERT TestOne VALUES(GETDATE())
INSERT TestOne VALUES(GETDATE())
INSERT TestOne VALUES(GETDATE())



Here are 6 ways to check for the current value

--1 @@IDENTITY
SELECT @@IDENTITY
--this returns 4


--2 DBCC CHECKIDENT
DBCC CHECKIDENT (TestOne, NORESEED)
after running DBCC CHECKIDENT the message returned is
Checking identity information: current identity value '4', current column value '4'.
DBCC execution completed. If DBCC printed error messages, contact your system administrator.


--3 MAX function
SELECT MAX(id)
FROM TestOne
you can also check with the MAX function but this is not recommended becuase you might get some other identity value that is not yours but from a different user

--4 TOP 1 and ORDER BY DESC
SELECT TOP 1 id
FROM TestOne
ORDER BY id DESC
--The same applies here as for the max function, this is not recommended


--5 IDENT_CURRENT
SELECT IDENT_CURRENT('TestOne')
--IDENT_CURRENT is another way to check


--6 SCOPE_IDENTITY
SELECT SCOPE_IDENTITY()
--This one is very similar to @@IDENTITY with one BIG difference (shown later)



--Let's add a trigger to the TestOne table
CREATE TRIGGER trTestOne ON [dbo].[TestOne]
FOR INSERT
AS
DECLARE
@CreditUserID INT

SELECT @CreditUserID = (SELECT ID FROM Inserted)

INSERT TestTwo VALUES(@CreditUserID,GETDATE())
GO

--Let's insert another row into the TestOne table
INSERT TestOne VALUES(GETDATE())

SELECT @@IDENTITY --1
SELECT @@IDENTITY,SCOPE_IDENTITY(),IDENT_CURRENT('TestOne') --1,5,5

Now why doesn't @@IDENTITY return 5 but 1?
This is because @@IDENTITY doesn't care about scope and returns the last identity value from all the statements, which in this case is from the code within the trigger trTestOne




So the bottom line is this: Always use SCOPE_IDENTITY() unless you DO need the last identity value regradless of scope (for example you need to know the identity from the table insert inside the trigger)

--Clean up this mess
DROP TABLE TestOne,TestTwo

Wednesday, July 26, 2006

@@ROWCOUNT And Trigger Woes

Here I go again linking to other people's posts instead of writing my own ;-(
Louis Davidson has a nice little entry on his SQL Doctor blog about Triggers and the @@ROWCOUNT function
You can check it out here

Tuesday, July 25, 2006

Integer Math In SQL Server

If you divide with 2 integers in SQL Server the result will be also an integer( 3/2 =1, 5/2 =2)
If you need the fractions then you need to convert/cast to decimal/numeric or multiply one of the integers by 1.0

Let's test this out

DECLARE @Value1 INT
DECLARE @Value2 INT

SELECT @Value1 =3,@Value2 =2

--here is where we get 1 instead of 1.5
SELECT @Value1/@Value2

--after using convert the result is correct
SELECT @Value1/CONVERT(DECIMAL(3,2),@Value2)

--after using cast the result is correct
SELECT @Value1/CAST(@Value2 AS DECIMAL(3,2))

--after multiplying with 1.0 the result is correct
SELECT @Value1/(@Value2*1.0)


--this won't work since the multiplication is done after we have the result
SELECT @Value1/@Value2*1.0

Return All the Rows From A Table Where At Least One Of The Columns IS NULL

Sometimes you inherit a table/database and you need to check if any of the rows in the table have columns with null values
I will show you a great timesaver to accomplish this task
First let's create the table

CREATE TABLE TestNulls (c1 INT,c2 INT, c3 INT,c4 INT,c5 INT, c6 INT)
INSERT TestNulls VALUES (1,1,1,1,1,1)
INSERT TestNulls VALUES (1,1,1,1,1,1)
INSERT TestNulls VALUES (1,1,1,1,1,1)
INSERT TestNulls VALUES (1,NULL,1,1,1,1)
INSERT TestNulls VALUES (1,1,1,1,1,1)
INSERT TestNulls VALUES (1,1,1,1,1,1)
INSERT TestNulls VALUES (1,1,1,NULL,1,1)

One way to check is to write an OR for every column
SELECT *
FROM TestNulls
WHERE C1 IS NULL
OR C2 IS NULL
OR C3 IS NULL
OR C4 IS NULL
OR C5 IS NULL
OR C6 IS NULL

A better (faster) way is to do it like this

SELECT *
FROM TestNulls
WHERE C1+C2+C3+C4+C5+C6 IS NULL

What if you have 50 columns, who wants to write them all out? Well not me.
No problem in Query Analyzer hit F8, this will bring up the Object Browser
In SQL Server Managment Studio use the Object Explorer
Navigate to the table click on the + and drag the whole column folder into the query window
You will see something like this c1, c2, c3, c4, c5, c6
Copy that open up notepad and paste it into notepad. Hit CTRL + H and in the Find box type a , and in the Replace box type a +, hit OK. After that do a CTRL + A and CTRL + C
Paste it back into Query Analyzer/SQL Server Managment Studio and you are done with that part

This will work everytime if all the columns are integers, what if you have varchars?
Let's test that out


CREATE TABLE TestNullsChar (c1 CHAR(1),c2 CHAR(1), c3 CHAR(1),c4 CHAR(1),c5 CHAR(1), c6 CHAR(1))
INSERT TestNullsChar VALUES (1,1,1,1,1,1)
INSERT TestNullsChar VALUES (1,1,1,1,1,1)
INSERT TestNullsChar VALUES (1,1,1,1,1,1)
INSERT TestNullsChar VALUES (1,NULL,1,1,1,1)
INSERT TestNullsChar VALUES (1,1,1,1,1,1)
INSERT TestNullsChar VALUES (1,1,1,1,1,1)
INSERT TestNullsChar VALUES (1,1,1,NULL,1,1)


No problem here (depending on your CONCAT_NULL_YIELDS_NULL setting)

SELECT *,C1+C2+C3+C4+C5+C6
FROM TestNullsChar
WHERE C1+C2+C3+C4+C5+C6 IS NULL

Let's set our CONCAT_NULL_YIELDS_NULL setting to off
SET CONCAT_NULL_YIELDS_NULL OFF

Run the query again

SELECT *,C1+C2+C3+C4+C5+C6
FROM TestNullsChar
WHERE C1+C2+C3+C4+C5+C6 IS NULL


As you can see nothing was returned so make sure that CONCAT_NULL_YIELDS_NULL is set to on before executing such a query because NULL + any other value will return NULL if CONCAT_NULL_YIELDS_NULL is set to on

SET CONCAT_NULL_YIELDS_NULL ON

Monday, July 24, 2006

Undocumented Options For UPDATE STATISTICS

Since I am still sleep deprived and can't think straight I have decided to post a link to a good article on the Tips, Tricks, and Advice from the SQL Server Query Optimization Team blog about some undocumented options that can be used with UPDATE STATISTICS

The link to that post is here: UPDATE STATISTICS Undocumented Options

Wednesday, July 19, 2006

A Father Again

I am a father again; my wife gave birth to 2 beautiful babies on Monday night. The boy is named Nicholas and the girl is name Catherine (yes just like Catherine the Great and Czar Nicholas II) I was working Monday and knew we were launching some new products this Wednesday. So I did all the work on staging on Monday and said to myself I will put it on the production box on Tuesday just in case my wife goes into labor. Well my wife didn’t feel that good on Monday and went in for a checkup and they decided to deliver the babies the same day. My son was with my wife so the person who I report to drove me to the hospital (I did not bring my keys to work) after that he drove me home so I could get some stuff and then he drove me back again (who else has a boss like that? ) A friend of ours came to the hospital to watch our son because my wife insisted that I had to be in the delivery room. My wife’s sisters were on vacation in Ocean City, they decided to visit that night and drove to the hospital. They arrived during the time that the babies were born. After the babies were born the sister took our son to our home and stayed there and I slept on the most comfortable (yeah right) hospital pull out chair. The sisters left on Tuesday and I went back home at 9PM. After putting my son to bed I logged in to my computer at work and worked for about an hour to move all the code over and to test that I didn’t mess up anything. My wife is still in the hospital but she will be coming home with the babies Friday morning. That’s it for now, next post will be SQL related again, probably sometime next week.

Below are some pictures, click on the pictures to get a bigger picture

Susan, Christian, Catherine and Nicholas


Nicholas


Catherine

Monday, July 17, 2006

Flat File Bulk Import Speed Comparison In SQL Server 2005

Mladen has compared the following 4 methods for importing a flat file in SQl Server 2005

1. BCP
2. Bulk Insert
3. OpenRowset with BULK option
4. SQL Server Integration Services - SSIS

Which do you think is the fastest? If you answered 1 or 2 you might be in for a surprise. Visit Mladen's blog to find out the answer

Three Ways To Get The Time Portion Of A Datetime Value

This popped up in a newsgroup today so I decided to do a little post about it. How do you get the time information only from a datetime value
These are at least 3 ways to do this

1) use the 3 different datepart functions and concatenate them
2) convert and grab the 8 right most characters
3) convert to varchar and use style 108

So here we go

--1 use the 3 different datepart functions and concatenate them
SELECT CONVERT(VARCHAR(2),DATEPART(HH,GETDATE())) + ':' +
CONVERT(VARCHAR(2),DATEPART(MI,GETDATE())) + ':' +
CONVERT(VARCHAR(2),DATEPART(S,GETDATE()))

--2 convert and grab the 8 right most characters
SELECT RIGHT(CONVERT(VARCHAR(22),GETDATE(),120),8)


--3 convert to varchar and use style 108
SELECT CONVERT(VARCHAR(12),GETDATE(),108)

Thursday, July 13, 2006

Visual Studio 2005 Team Edition for Database Professionals Interview On Channel 9

Channel 9 has a cool webcast with several people talking about Visual Studio 2005 Team Edition for Database Professionals (Data Dude)

From the site: "Cameron Skinner, Gert Drapers, Robert Merriman, Thomas Murphy, and Matt Nunn sat down me with to discuss a new edition to the Visual Studio Team product family: Visual Studio 2005 Team Edition for Database Professionals. We talked about why this product was created and Cameron (the product unit manager) gave a walk through of the current CTP release. Targeted for delivery at the end of the year, you can grab the CTP bits and learn more at the team’s website: http://msdn.microsoft.com/vstudio/teamsystem/dbpro/"


You can download the webcast here (http://channel9.msdn.com/Showpost.aspx?postid=212797) or click on the picture

Visual Studio 2005 Team Edition for Database Professionals CTP 4 Available

Visual Studio 2005 Team Edition for Database Professionals (Data Dude) CTP 4 is available for download

Brief Description
Community Technology Preview (CTP) 4
Tools for building SQL databases in a managed project environment with support for versioning, deployment, unit testing, refactoring, and off-line SQL development.

Overview
Tools for building SQL databases in a managed project environment with support for versioning, deployment, unit testing, refactoring, and off-line SQL development. This release is focused on completing a full functionality scenario for SQL 2000 and includes an updated project system with reverse engineering, a new SQL Query Editor, Schema and Data Compare, Data Generation, Unit Testing, and Rename Refactoring. This release does not have functionality changes from CTP3, only stability work and the new project UI.

Get it here

Wednesday, July 12, 2006

SqlBlog

Peter DeBetta and Adam Machanic have created SqlBlog.com. It has a very nice list of SQL Server blogs and if you click on roller you will see a about 40 of them displayed (mine is missing ;-( )
So what are you waiting for? Sign up here, I already did

Tuesday, July 11, 2006

Non Updating Update Performance Improvement In SQL Server 2005

When you do a non updating update like the one below

UPDATE t
SET i = 1
WHERE i=1



In SQL 2000, the lack of the optimization leads to updating the nonclustered index even if the value is not changing.

In the SQL 2005 plan, it is possible to appreciate
- a “Compute Scalar” operator that compares the current value and new value of the column being modified
- a new filter operator that on a row by row basis will determine whether the nonclustered index is being affected or not
- the fact that nonclustered index maintenance is now bypassed

Read the whole article (including statistics profile output screenshot) at the Tips, Tricks, and Advice from the SQL Server Query Optimization Team blog

Monday, July 10, 2006

Setting Identity Value Back To 1 After Deleting All Rows From A Table

Sometimes you want the identity value to start from 1 again after you delete all the rows from a table
There are 2 ways to accomplish that
1 Use the TRUNCATE TABLE command instead of DELETE
2 If you do use DELETE then run DBCC CHECKIDENT and reseed the table

Here is some code to explain what I mean

CREATE TABLE TestValues (Id INT identity,DateVale DATETIME)
INSERT INTO TestValues VALUES(GETDATE())
INSERT INTO TestValues VALUES(GETDATE())
INSERT INTO TestValues VALUES(GETDATE())
INSERT INTO TestValues VALUES(GETDATE())
INSERT INTO TestValues VALUES(GETDATE())

SELECT *
FROM TestValues

-- This is the output
-- 1 2006-07-10 12:31:17.050
-- 2 2006-07-10 12:31:17.050
-- 3 2006-07-10 12:31:17.050
-- 4 2006-07-10 12:31:17.050
-- 5 2006-07-10 12:31:17.050



DELETE TestValues

INSERT INTO TestValues VALUES(GETDATE())

SELECT *
FROM TestValues

-- Output
-- 6 2006-07-10 12:31:29.143


TRUNCATE TABLE TestValues

INSERT INTO TestValues VALUES(GETDATE())

SELECT *
FROM TestValues

-- Output
-- 1 2006-07-10 12:31:38.317


INSERT INTO TestValues VALUES(GETDATE())

DELETE TestValues

DBCC CHECKIDENT(TestValues,RESEED,0)
-- Message
-- Checking identity information: current identity value '2', current column value '0'.
-- DBCC execution completed. If DBCC printed error messages, contact your system administrator.


INSERT INTO TestValues VALUES(GETDATE())

SELECT *
FROM TestValues

-- Output
-- 1 2006-07-10 12:31:52.503


DROP TABLE TestValues

Building SQL Server Integration Services Packages

Learn how easy it is to build and debug your first SQL Server Integration Services (SSIS) Package. Donald Farmer introduces the development environment and the basic concepts of packages and walks through the creation and testing of a simple data flow. See why users describe SSIS as the most productive environment for data integration.

Watch the video at MSDN TV

Get the files here

Wednesday, July 05, 2006

Designing Effective Aggregations In Analysis Services 2005

If you're into Analysis Services, Integration Services, Reporting Services and BI in general - you'll want to check out Elizabeth Vitt's new blog I found the link to her blog on Kimberly Tripps blog

In her first entry, Elizabeth Vitt highlights the design components that determine which attributes are considered for aggregation:


Aggregation Definition
Attribute Relationships
Aggregation Candidates
Aggregation Usage

Read the article here: Influencing Aggregation Candidates