Tuesday, January 24, 2012

Ideal SQL Query For Handling Error & Transcation in MS SQL



BEGIN TRY

BEGIN TRAN

--put queries here

COMMIT;

END TRY
BEGIN CATCH
IF @@TRANCOUNT>0
BEGIN
SELECT @@ERROR,ERROR_MESSAGE ( )
ROLLBACK;
END
END CATCH


any one can use above code to manage transaction as well as error handling in sql server.

Tuesday, May 10, 2011

Capitalize the First Character of Each Word using queries

I have found t-sql challenges question where they were asked to provide a solution of oracle INITCAP function in sql.

I was started to find the solution of INITCAP function & found lots of user defined function.

But here I have shared a solution where queries doesn't use any user defined function as well as any variable.

--First of all create one temp table    
DECLARE @t TABLE(ID INT IDENTITY, Data VARCHAR(1000))

--insert the sample data
INSERT INTO @t 
SELECT 'yogesh bhadauriya' 


--use the Common table expression
;with CTE AS
(

    --anchor query where we are finding the first character for replacing with uppercase
    SELECT Cast(STUFF(LOWER(Data), 1, 1, UPPER(LEFT(Data, 1))) As varchar(max)) As Data,
    PATINDEX('%[^a-zA-Z][a-z]%', Data COLLATE Latin1_General_Bin) As Position,ID,Data As OldData
    From @t
    
    Union All
    
   --recursive query
    SELECT Cast(STUFF(CTE.Data, Position, 2, UPPER(SUBSTRING(CTE.Data, CTE.Position, 2))) As Varchar(max)) as Data,
    Cast(PATINDEX('%[^a-zA-Z][a-z]%', CTE.Data COLLATE Latin1_General_Bin) As Int) As Position,CTE.ID,T.Data As OldData
    From @t As T
    Inner Join CTE On CTE.ID=CTE.ID
    where CTE.Data is not null
)
Select ID,OldData As Data,Data As Results
From CTE
where Data is not null
and Position=0


---output
--Id Data    Results
--1 yogesh bhadauriya Yogesh Bhadauriya
    

Friday, April 29, 2011

User defined function to convert the number into charcter

How to convert the numeric value into its specific character? If this is your question, than create the following User Defined Function into your database.


--create the following function 
CREATE FUNCTION dbo.udf_Num_ToWords 
(
 @Number Numeric (38, 0) -- Input number with as many as 18 digits

) RETURNS VARCHAR(8000) 
/*
* Converts a integer number as large as 34 digits into the 
* equivalent words.  The first letter is capitalized.
*
* Attribution: Based on NumberToWords by Srinivas Sampath
*        as revised by Nick Barclay
*
* Example:
select dbo.udf_Num_ToWords (1234567890) + CHAR(10)
      +  dbo.udf_Num_ToWords (0) + CHAR(10)
      +  dbo.udf_Num_ToWords (123) + CHAR(10)
select dbo.udf_Num_ToWords(76543210987654321098765432109876543210)
 
DECLARE @i numeric (38,0)
SET @i = 0
WHILE @I <= 1000 BEGIN 
    PRINT convert (char(5), @i)  
            + convert(varchar(255), dbo.udf_Num_ToWords(@i)) 
    SET @I  = @i + 1 
END
*
* Published as the T-SQL UDF of the Week Vol 2 #9 2/17/03
****************************************************************/
AS BEGIN

DECLARE @inputNumber VARCHAR(38)
DECLARE @NumbersTable TABLE (number CHAR(2), word VARCHAR(10))
DECLARE @outputString VARCHAR(8000)
DECLARE @length INT
DECLARE @counter INT
DECLARE @loops INT
DECLARE @position INT
DECLARE @chunk CHAR(3) -- for chunks of 3 numbers
DECLARE @tensones CHAR(2)
DECLARE @hundreds CHAR(1)
DECLARE @tens CHAR(1)
DECLARE @ones CHAR(1)

IF @Number = 0 Return 'Zero'

-- initialize the variables
SELECT @inputNumber = CONVERT(varchar(38), @Number)
     , @outputString = ''
     , @counter = 1
SELECT @length   = LEN(@inputNumber)
     , @position = LEN(@inputNumber) - 2
     , @loops    = LEN(@inputNumber)/3

-- make sure there is an extra loop added for the remaining numbers
IF LEN(@inputNumber) % 3 <> 0 SET @loops = @loops + 1

-- insert data for the numbers and words
INSERT INTO @NumbersTable   SELECT '00', ''
    UNION ALL SELECT '01', 'one'      UNION ALL SELECT '02', 'two'
    UNION ALL SELECT '03', 'three'    UNION ALL SELECT '04', 'four'
    UNION ALL SELECT '05', 'five'     UNION ALL SELECT '06', 'six'
    UNION ALL SELECT '07', 'seven'    UNION ALL SELECT '08', 'eight'
    UNION ALL SELECT '09', 'nine'     UNION ALL SELECT '10', 'ten'
    UNION ALL SELECT '11', 'eleven'   UNION ALL SELECT '12', 'twelve'
    UNION ALL SELECT '13', 'thirteen' UNION ALL SELECT '14', 'fourteen'
    UNION ALL SELECT '15', 'fifteen'  UNION ALL SELECT '16', 'sixteen'
    UNION ALL SELECT '17', 'seventeen' UNION ALL SELECT '18', 'eighteen'
    UNION ALL SELECT '19', 'nineteen' UNION ALL SELECT '20', 'twenty'
    UNION ALL SELECT '30', 'thirty'   UNION ALL SELECT '40', 'forty'
    UNION ALL SELECT '50', 'fifty'    UNION ALL SELECT '60', 'sixty'
    UNION ALL SELECT '70', 'seventy'  UNION ALL SELECT '80', 'eighty'
    UNION ALL SELECT '90', 'ninety'   

WHILE @counter <= @loops BEGIN

 -- get chunks of 3 numbers at a time, padded with leading zeros
 SET @chunk = RIGHT('000' + SUBSTRING(@inputNumber, @position, 3), 3)

 IF @chunk <> '000' BEGIN
  SELECT @tensones = SUBSTRING(@chunk, 2, 2)
       , @hundreds = SUBSTRING(@chunk, 1, 1)
       , @tens = SUBSTRING(@chunk, 2, 1)
       , @ones = SUBSTRING(@chunk, 3, 1)

  -- If twenty or less, use the word directly from @NumbersTable
  IF CONVERT(INT, @tensones) <= 20 OR @Ones='0' BEGIN
   SET @outputString = (SELECT word 
                                      FROM @NumbersTable 
                                      WHERE @tensones = number)
                   + CASE @counter WHEN 1 THEN '' -- No name
                       WHEN 2 THEN ' thousand ' WHEN 3 THEN ' thousand '
                       WHEN 4 THEN ' billion '  WHEN 5 THEN ' trillion '
                       WHEN 6 THEN ' quadrillion ' WHEN 7 THEN ' quintillion '
                       WHEN 8 THEN ' sextillion '  WHEN 9 THEN ' septillion '
                       WHEN 10 THEN ' octillion '  WHEN 11 THEN ' nonillion '
                       WHEN 12 THEN ' decillion '  WHEN 13 THEN ' undecillion '
                       ELSE '' END
                               + @outputString
      END
   ELSE BEGIN -- break down the ones and the tens separately

             SET @outputString = ' ' 
                            + (SELECT word 
                                    FROM @NumbersTable 
                                    WHERE @tens + '0' = number)
              + '-'
                             + (SELECT word 
                                    FROM @NumbersTable 
                                    WHERE '0'+ @ones = number)
                   + CASE @counter WHEN 1 THEN '' -- No name
                       WHEN 2 THEN ' thousand ' WHEN 3 THEN ' million '
                       WHEN 4 THEN ' billion '  WHEN 5 THEN ' trillion '
                       WHEN 6 THEN ' quadrillion ' WHEN 7 THEN ' quintillion '
                       WHEN 8 THEN ' sextillion '  WHEN 9 THEN ' septillion '
                       WHEN 10 THEN ' octillion '  WHEN 11 THEN ' nonillion '
                       WHEN 12 THEN ' decillion '   WHEN 13 THEN ' undecillion '
                       ELSE '' END
                            + @outputString
  END

  -- now get the hundreds
  IF @hundreds <> '0' BEGIN
   SET @outputString  = (SELECT word 
                                      FROM @NumbersTable 
                                      WHERE '0' + @hundreds = number)
                 + ' hundred ' 
                                + @outputString
  END
 END

 SELECT @counter = @counter + 1
      , @position = @position - 3

END

-- Remove any double spaces
SET @outputString = LTRIM(RTRIM(REPLACE(@outputString, '  ', ' ')))
SET @outputstring = UPPER(LEFT(@outputstring, 1)) + SUBSTRING(@outputstring, 2, 8000)


RETURN @outputString -- return the result
END





To check the output of the above User Defined Function, execute the below query into your database.

--now run the following query to get the result

select dbo.udf_Num_ToWords(5800000)

--output
--Five thousand eight hundred thousand


--
Happy Coding to All

Recursive Query To Find All Child Node in SQL server 2005+

How to find the all child of particular parent using the Recursive Query
--here we are creating the table variable so that we can insert some dummy record
Declare @Table Table
(
 TId int,
 ParentId int,
 Name varchar(10)
)
--inserting some records
--using the union all to insert more than one records
insert into @Table
Select 1,NULL,'ICT'
Union All
Select 2,1,'ICT-M1'
Union All
Select 4,1,'ICT-M2'
Union All
Select 7,2,'ICT-M1U1'
Union All
Select 8,2,'ICT-M1U2'
Union All
Select 9,4,'ICT-M2U1'
Union All
Select 10,4,'ICT-M2U2'
Union All
Select 11,7,'ICT-M1U1P1'
Union All
Select 12,7,'ICT-M1U1P2'
Union All
Select 13,8,'ICT-M1U2P1'
Union All
Select 14,8,'ICT-M1U2P2'
Union All
Select 15,9,'ICT-M2U1P1'
Union All
Select 16,9,'ICT-M2U1P2'
Union All
Select 17,10,'ICT-M2U2P1'
Union All
Select 18,10,'ICT-M2U2P2'

--variable to hold data
Declare @ChildNode varchar(1000)
Set @ChildNode='';

 
--use the standard recursive query
;with [CTE] as 
(
 --anchor query where we are finding the all parents
    select TId,ParentId,Name,CAST(ISNULL(CAST(ParentId as varchar(10)),'0') As Varchar(100)) As ChildNode
    from @Table c where c.[ParentId] is null

    union all

 --recursive query where we are finding the all child according to anchor query parent
    select c.TId,c.ParentId,c.Name,
    CAST( p.ChildNode +','+cast(c.TId as varchar(10) ) As Varchar(100)) As ChildNode
    from [CTE] p, @Table c where c.[ParentId] = p.[TId]    
)
--select the child node as per the id
--Assigin the all Ids into  one variable
select @ChildNode=@ChildNode+','+Cast(TId as varchar(10))
from [CTE]
Cross Apply
dbo.Split(ChildNode,',')
where items=9
order by TId

select SUBSTRING(@ChildNode,2,LEN(@ChildNode))

--outpu
--2,7,8,11,12,13,14


----
--create the tabular function to split the multi valued into table
Create FUNCTION [dbo].[Split](@String varchar(8000), @Delimiter char(1))     
returns @temptable TABLE (items varchar(8000))     
as     
begin     
 declare @idx int     
 declare @slice varchar(8000)     
    
 select @idx = 1     
  if len(@String)<1 or @String is null  return     
    
 while @idx!= 0     
 begin     
  set @idx = charindex(@Delimiter,@String)     
  if @idx!=0     
   set @slice = left(@String,@idx - 1)     
  else     
   set @slice = @String     
  
  if(len(@slice)>0)
   insert into @temptable(Items) values(@slice)     

  set @String = right(@String,len(@String) - @idx)     
  if len(@String) = 0 break     
 end 
return     
end


--
Happy Coding

Saturday, April 16, 2011

How to insert multiple record using xml in sql server 2005+

How to insert the multiple record using the xml into sql server, please check the following example.


DECLARE @idoc int
DECLARE @doc varchar(max)
SET @doc ='
<ROOT>
 <Trans TransId="1" Add="false" Edit="true" Delete="true" View="true" Block="false">   
 </Trans>
 <Trans TransId="2" Add="1" Edit="1" Delete="1" View="1" Block="0">   
 </Trans>
</ROOT>'

--Create an internal representation of the XML document.
EXEC sp_xml_preparedocument @idoc OUTPUT, @doc
-- Execute a SELECT statement that uses the OPENXML rowset provider.


SELECT *
Into #TempTable 
FROM OPENXML(@idoc, '/ROOT/Trans',1)
WITH 
( 
 TransId  varchar(10),
    [Add] bit,
    Edit bit,
    [Delete] bit,
    [View] bit,
    Block bit
)

Select * From #TempTable

drop table #TempTable

--
happy coding 

Tuesday, December 14, 2010

Performance Comparision Between For, While and Foreach Loop

Today I was posted one post regarding list to datatable conversion. But one of the my senior told me that try to avoid foreach loop.

So that I was googled/binged regarding the same.

Here I am showing you some figure for the same here.

Using the System.Diagnostics.Stopwatch class I ran some tests. 100,000 iterations in a for loop that did nothing inside took me 0.0003745 seconds. This was the code for the loop:
 
for (int i = 0; i < 100000; i++) ;

The while loop resulted 0.0003641 seconds, which is pretty much the same as the for loop. Here is the code I used:

int i=0;
while (i < 100000)
 i++;

 
The foreach loop has a slightly different purpose. It is meant for itterating through some collection that implements IEnumerable. It's performance is much slower, my test resulted in 0.0009076 seconds with this code:

int[] test = new int[100000];
foreach (int i in test) ;

foreach creates an instance of an enumerator (returned from GetEnumerator) and that enumerator also keeps state throughout the course of the foreach loop. It then repeatedly calls for the Next() object on the enumerator and runs your code for each object it returns.


So, it seems that 'while' is the fastest looping technique among the three available techniques in C#,  for a given processing within the loop. Right?

It varies. "while" and "for" have pretty much the same results. While had a slight advantage, but not one that great.


Im not an expert, but I have the feeling that a while and for loop, once compiled to MSLI, probably are both the exact same thing.

And I wouldn't be surprised if foreach was faster when iterating through objects, since the optimizer can "expect" whats going to happen... So I'd go with foreach being faster if you can use it, and the 2 others being the same thing, if foreach isn't applicable.

My logic here is that if you're looping through a collection, and you use a for loop for example, you're going to have to use the index of the collection, which depending on implementation, could be a minor performance hit to "seek" the object, as opposed to going through a well written iterator.

Just an example.So really: I wouldn't care too much about performance of these loops. This isn't C/C++, and when you compile, its not native code (at first). So it is safe to assume that the solution that seems the most efficient "logically" will be so in practice.

For each is slower for a number of reasons. One is it is using the IEnumerable interface, which requires some casting (assuming you aren't using a generic collection). My test above seemed to go along with that as well. A simple for/while/do loop is pretty much as simple as it gets.

As for the MSIL code, let's take a look. This is the MSIL for the for loop:

IL_0000:  nop
  IL_0001:  ldc.i4.0
  IL_0002:  stloc.0
  IL_0003:  br.s       IL_0009
  IL_0005:  ldloc.0
  IL_0006:  ldc.i4.1
  IL_0007:  add
  IL_0008:  stloc.0
  IL_0009:  ldloc.0
  IL_000a:  ldc.i4     0x186a0
  IL_000f:  clt
  IL_0011:  stloc.1
  IL_0012:  ldloc.1
  IL_0013:  brtrue.s   IL_0005

And here it is for the while loop:

IL_0000:  nop
  IL_0001:  ldc.i4.0
  IL_0002:  stloc.0
  IL_0003:  br.s       IL_0009
  IL_0005:  ldloc.0
  IL_0006:  ldc.i4.1
  IL_0007:  add
  IL_0008:  stloc.0
  IL_0009:  ldloc.0
  IL_000a:  ldc.i4     0x186a0
  IL_000f:  clt
  IL_0011:  stloc.1
  IL_0012:  ldloc.1
  IL_0013:  brtrue.s   IL_0005

So yes, they are exactly identical.

Although very handy, C#'s foreach statement is actually quite dangerous. In fact, I may swear off its use entirely. Why? Two reasons: (1) performance, and (2) predictability.

Performance

Iterating through a collection using foreach is slower than with for. I can't remember where I first learned that, perhaps in Patterns & Practices: Improving .Net Application Performance. Maybe it was from personal experience. How much slower? Well, I suppose that depends on your particular circumstances. Here are a few interesting references:

Predictability

I was looking at the C# Reference entry for foreach today and noticed this for the first time (italics added by me):

The foreach statement is used to iterate through the collection to get the desired information, but should not be used to change the contents of the collection to avoid unpredictable side effects.
What's that all about? Let's take this as an example:
foreach(MyClass myObj in List)
Looking deeper into the C# Language Specification... the iteration variable is supposed to be read-only, though apparently that doesn't stop you from updating a property of an object. Thus for instance it would be illegal to assign a new value to myObj, but not to assign a new value to myObj.MyProperty.
And that's all I can find. Why are there unpredictable side effects? I don't know. But seems best to heed Microsoft's warning.

Conclusion

Some argue that you shouldn't code for performance from the beginning, and therefore go ahead and use foreach whenever you want so long as you don't update the values. In my experience that's hogwash — most of the code I work on goes into environments where performance is extremely important. Besides, writing a for statement requires very little extra coding compared to a foreach statement. Furthermore, if you have a lot going on inside your iteration block, it can be easy to forget and accidentally update the iteration variable inside a foreach loop. Thus do I conclude: just avoid foreach altogether.

Honestly the academically correct answer is "It's Irrelevant".  You cant optimize performance by somehow picking "the best loop".  If you're doing performance optimizations you should start with a Big O analysis of your algorithms and Profiling.  You'd be amazed at how much faster keeping around a dictionary for lookups or using a sorted list + binary search is than looping over a list every time you need to find an object.

Doing so will give you unnoticeable performance increases at the cost of maintainability and programmer time and the effort you're putting in to such small performance boosts is better spent on a proper design and implementation.   Let's take a real world example of finding duplicates in a list to illustrate my point:
first let's assume that comparing two items is O(1).  We can implement this any number of ways, two of which are:

A) use nested loops with i on the outer loop and j on the inner loop, when list[i] == list[j] push the pair onto a list of duplicates.
B) copy the list to tmpList.  quicksort tmpList.  iterate over the list with i, when list[i] == list[i+1] push the pair onto the list of duplicates

No amount of optimization can change the fact that A runs in O(n^2) while B runs in O(2*n+n*log(n)).

Convert List to DataTable

One day I confused to convert the List collection to datatable object. I was wondering how to iterate the columns wise and rows wise in list to make some calculation.

I got the solution which was posted below:

public DataTable ListToDataTable(IEnumerable list)
{
    var dt = new DataTable();

    foreach (var info in typeof(T).GetProperties())
    {
        dt.Columns.Add(new DataColumn(info.Name, info.PropertyType));
    }
    foreach (var t in list)
    {
        var row = dt.NewRow();
        foreach (var info in typeof(T).GetProperties())
        {
            row[info.Name] = info.GetValue(t, null);
        }
        dt.Rows.Add(row);
    }
    return dt;
}

--
Happy coding.

Ideal SQL Query For Handling Error & Transcation in MS SQL

BEGIN TRY BEGIN TRAN --put queries here COMMIT; END TRY BEGIN CATCH IF @@TRANCOUNT>0 BEGIN SELECT @@ERROR,ERRO...