Search Results

Search found 6 results on 1 pages for 'asanka'.

Page 1/1 | 1 

  • Importing data from text file to specific columns using BULK INSERT

    - by Dinesh Asanka
    Bulk insert is much faster than using other techniques such as  SSIS. However, when you are using bulk insert you can’t insert to specific columns. If, for example, there are five columns in a table you should have five values for each record in the text file you are importing from. This is an issue when you are expecting default values to be inserted into tables. Let us say you have table as below: In this table, you are expecting ID, Status and CreatedDate to be updated automatically, so your text file may only have   FirstName  LastName  values as below: Dinesh,Asanka Saman,Liyanage Ruwan,Silva Susantha,Bathige Jude,Peires Sanjeewa,Jayawickrama If you use bulk insert to this table like follows, You will be returned an error: Bulk load data conversion error (type mismatch or invalid character for the specified codepage) for row 1, column 1 (ID). To avoid this you will need to create a view with the columns you are expecting to fill and use bulk insert against it. If you check the table now, you will see table with values in the text file and the default values.

    Read the article

  • Whether to use UNION or OR in SQL Server Queries

    - by Dinesh Asanka
    Recently I came across with an article on DB2 about using Union instead of OR. So I thought of carrying out a research on SQL Server on what scenarios UNION is optimal in and which scenarios OR would be best. I will analyze this with a few scenarios using samples taken  from the AdventureWorks database Sales.SalesOrderDetail table. Scenario 1: Selecting all columns So we are going to select all columns and you have a non-clustered index on the ProductID column. --Query 1 : OR SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 714 OR ProductID =709 OR ProductID =998 OR ProductID =875 OR ProductID =976 OR ProductID =874 --Query 2 : UNION SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 714 UNION SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 709 UNION SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 998 UNION SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 875 UNION SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 976 UNION SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 874 So query 1 is using OR and the later is using UNION. Let us analyze the execution plans for these queries. Query 1 Query 2 As expected Query 1 will use Clustered Index Scan but Query 2, uses all sorts of things. In this case, since it is using multiple CPUs you might have CX_PACKET waits as well. Let’s look at the profiler results for these two queries: CPU Reads Duration Row Counts OR 78 1252 389 3854 UNION 250 7495 660 3854 You can see from the above table the UNION query is not performing well as the  OR query though both are retuning same no of rows (3854).These results indicate that, for the above scenario UNION should be used. Scenario 2: Non-Clustered and Clustered Index Columns only --Query 1 : OR SELECT ProductID,SalesOrderID, SalesOrderDetailID FROM Sales.SalesOrderDetail WHERE ProductID = 714 OR ProductID =709 OR ProductID =998 OR ProductID =875 OR ProductID =976 OR ProductID =874 GO --Query 2 : UNION SELECT ProductID,SalesOrderID, SalesOrderDetailID FROM Sales.SalesOrderDetail WHERE ProductID = 714 UNION SELECT ProductID,SalesOrderID, SalesOrderDetailID FROM Sales.SalesOrderDetail WHERE ProductID = 709 UNION SELECT ProductID,SalesOrderID, SalesOrderDetailID FROM Sales.SalesOrderDetail WHERE ProductID = 998 UNION SELECT ProductID,SalesOrderID, SalesOrderDetailID FROM Sales.SalesOrderDetail WHERE ProductID = 875 UNION SELECT ProductID,SalesOrderID, SalesOrderDetailID FROM Sales.SalesOrderDetail WHERE ProductID = 976 UNION SELECT ProductID,SalesOrderID, SalesOrderDetailID FROM Sales.SalesOrderDetail WHERE ProductID = 874 GO So this time, we will be selecting only index columns, which means these queries will avoid a data page lookup. As in the previous case we will analyze the execution plans: Query 1 Query 2 Again, Query 2 is more complex than Query 1. Let us look at the profile analysis: CPU Reads Duration Row Counts OR 0 24 208 3854 UNION 0 38 193 3854 In this analyzis, there is only slight difference between OR and UNION. Scenario 3: Selecting all columns for different fields Up to now, we were using only one column (ProductID) in the where clause.  What if we have two columns for where clauses and let us assume both are covered by non-clustered indexes? --Query 1 : OR SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 714 OR CarrierTrackingNumber LIKE 'D0B8%' --Query 2 : UNION SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 714 UNION SELECT * FROM Sales.SalesOrderDetail WHERE CarrierTrackingNumber  LIKE 'D0B8%' Query 1 Query 2: As we can see, the query plan for the second query has improved. Let us see the profiler results. CPU Reads Duration Row Counts OR 47 1278 443 1228 UNION 31 1334 400 1228 So in this case too, there is little difference between OR and UNION. Scenario 4: Selecting Clustered index columns for different fields Now let us go only with clustered indexes: --Query 1 : OR SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 714 OR CarrierTrackingNumber LIKE 'D0B8%' --Query 2 : UNION SELECT * FROM Sales.SalesOrderDetail WHERE ProductID = 714 UNION SELECT * FROM Sales.SalesOrderDetail WHERE CarrierTrackingNumber  LIKE 'D0B8%' Query 1 Query 2 Now both execution plans are almost identical except is an additional Stream Aggregate is used in the first query. This means UNION has advantage over OR in this scenario. Let us see profiler results for these queries again. CPU Reads Duration Row Counts OR 0 319 366 1228 UNION 0 50 193 1228 Now see the differences, in this scenario UNION has somewhat of an advantage over OR. Conclusion Using UNION or OR depends on the scenario you are faced with. So you need to do your analyzing before selecting the appropriate method. Also, above the four scenarios are not all an exhaustive list of scenarios, I selected those for the broad description purposes only.

    Read the article

  • Altering a Column Which has a Default Constraint

    - by Dinesh Asanka
    Setting up a default column is a common task for  developers.  But, are we naming those default constraints explicitly? In the below  table creation, for the column, sys_DateTime the default value Getdate() will be allocated. CREATE TABLE SampleTable (ID int identity(1,1), Sys_DateTime Datetime DEFAULT getdate() ) We can check the relevant information from the system catalogs from following query. SELECT sc.name TableName, dc.name DefaultName, dc.definition, OBJECT_NAME(dc.parent_object_id) TableName, dc.is_system_named  FROM sys.default_constraints dc INNER JOIN sys.columns sc ON dc.parent_object_id = sc.object_id AND dc.parent_column_id = sc.column_id and results would be: Most of the above columns are self-explanatory. The last column, is_system_named, is to identify whether the default name was given by the system. As you know, in the above case, since we didn’t provide  any default name, the  system will generate a default name for you. But the problem with these names is that they can differ from environment to environment.  If example if I create this table in different table the default name could be DF__SampleTab__Sys_D__7E6CC920 Now let us create another default and explicitly name it: CREATE TABLE SampleTable2 (ID int identity(1,1), Sys_DateTime Datetime )   ALTER TABLE SampleTable2 ADD CONSTRAINT DF_sys_DateTime_Getdate DEFAULT( Getdate()) FOR Sys_DateTime If we run the previous query again we will be returned the below output. And you can see that last created default name has 0 for is_system_named. Now let us say I want to change the data type of the sys_DateTime column to something else: ALTER TABLE SampleTable2 ALTER COLUMN Sys_DateTime Date This will generate the below error: Msg 5074, Level 16, State 1, Line 1 The object ‘DF_sys_DateTime_Getdate’ is dependent on column ‘Sys_DateTime’. Msg 4922, Level 16, State 9, Line 1 ALTER TABLE ALTER COLUMN Sys_DateTime failed because one or more objects access this column. This means, you need to drop the default constraint before altering it: ALTER TABLE [dbo].[SampleTable2] DROP CONSTRAINT [DF_sys_DateTime_Getdate] ALTER TABLE SampleTable2 ALTER COLUMN Sys_DateTime Date   ALTER TABLE [dbo].[SampleTable2] ADD CONSTRAINT [DF_sys_DateTime_Getdate] DEFAULT (getdate()) FOR [Sys_DateTime] If you have a system named default constraint that can differ from environment to environment and so you cannot drop it as before, you can use the below code template: DECLARE @defaultname VARCHAR(255) DECLARE @executesql VARCHAR(1000)   SELECT @defaultname = dc.name FROM sys.default_constraints dc INNER JOIN sys.columns sc ON dc.parent_object_id = sc.object_id AND dc.parent_column_id = sc.column_id WHERE OBJECT_NAME (parent_object_id) = 'SampleTable' AND sc.name ='Sys_DateTime' SET @executesql = 'ALTER TABLE SampleTable DROP CONSTRAINT ' + @defaultname EXEC( @executesql) ALTER TABLE SampleTable ALTER COLUMN Sys_DateTime Date ALTER TABLE [dbo].[SampleTable] ADD DEFAULT (Getdate()) FOR [Sys_DateTime]

    Read the article

  • Adding Column to a SQL Server Table

    - by Dinesh Asanka
    Adding a column to a table is  common task for  DBAs. You can add a column to a table which is a nullable column or which has default values. But are these two operations are similar internally and which method is optimal? Let us start this with an example. I created a database and a table using following script: USE master Go --Drop Database if exists IF EXISTS (SELECT 1 FROM SYS.databases WHERE name = 'AddColumn') DROP DATABASE AddColumn --Create the database CREATE DATABASE AddColumn GO USE AddColumn GO --Drop the table if exists IF EXISTS ( SELECT 1 FROM sys.tables WHERE Name = 'ExistingTable') DROP TABLE ExistingTable GO --Create the table CREATE TABLE ExistingTable (ID BIGINT IDENTITY(1,1) PRIMARY KEY CLUSTERED, DateTime1 DATETIME DEFAULT GETDATE(), DateTime2 DATETIME DEFAULT GETDATE(), DateTime3 DATETIME DEFAULT GETDATE(), DateTime4 DATETIME DEFAULT GETDATE(), Gendar CHAR(1) DEFAULT 'M', STATUS1 CHAR(1) DEFAULT 'Y' ) GO -- Insert 100,000 records with defaults records INSERT INTO ExistingTable DEFAULT VALUES GO 100000 Before adding a Column Before adding a column let us look at some of the details of the database. DBCC IND (AddColumn,ExistingTable,1) By running the above query, you will see 637 pages for the created table. Adding a Column You can add a column to the table with following statement. ALTER TABLE ExistingTable Add NewColumn INT NULL Above will add a column with a null value for the existing records. Alternatively you could add a column with default values. ALTER TABLE ExistingTable Add NewColumn INT NOT NULL DEFAULT 1 The above statement will add a column with a 1 value to the existing records. In the below table I measured the performance difference between above two statements. Parameter Nullable Column Default Value CPU 31 702 Duration 129 ms 6653 ms Reads 38 116,397 Writes 6 1329 Row Count 0 100000 If you look at the RowCount parameter, you can clearly see the difference. Though column is added in the first case, none of the rows are affected while in the second case all the rows are updated. That is the reason, why it has taken more duration and CPU to add column with Default value. We can verify this by several methods. Number of Pages The number of data pages can be obtained by using DBCC IND command. Though, this an undocumented dbcc command, many experts are ok to use this command in production. However, since there is no official word from Microsoft, use this “at your own risk”. DBCC IND (AddColumn,ExistingTable,1) Before Adding the Columns 637 Adding a Column with NULL 637 Adding a column with DEFAULT value 1270 This clearly shows that pages are physically modified. Please note, a high value indicated in the Adding a column with DEFAULT value  column is also a result of page splits. Continues…

    Read the article

  • Compile OpenSSL from source using Xilinx EDK

    - by Asanka
    I want to use openSSL functions in my xilinx C++ project. So i downloaded the source and tried to compile using the makefile. But unfortunately the linux box I was provided did not contained the required software (perl5, make, etc) to compile the source using makefile. So as the second option i tried to compile the source as a shared library. So i made a new project using existing openSSL source. But all the attempts were failed since it gave lots of errors including missing headers, unknown defines etc. If someone knows how to compile openssl as a shared library using Xilinx toolchains please help me. It would be really helpful if someone can provide the complete project which can be opened using Xilinx EDK. If better alternative exists to use openssl with xilinx toolchains please explain that also. Thanks in advance....

    Read the article

1