SQL Server cursor -

What is cursor in SQL Server?

• A Cursor is a database object that represents a result set and is used to manipulate data row by row.
• When a cursor is opened, it is positioned on a row and that row is available for processing.
• SQL Server supports three types of cursor namely Transact-SQL server cursor, API server cursor, and client cursor.
• Transact-SQL Server cursors use Transact-SQL statements and are declared using DECLARE CURSOR statement.
• Transact-SQL Server cursors can be used in Transact-SQL scripts, stored procedures, and triggers.
• Transact-SQL cursors are implemented on the server.
• You can fetch only one row at a time in Transact-SQL Server cursors.
• You can use FETCH statements with Transact-SQL cursors to retrieve rows from a cursor’s result set.
• API server cursors support the API cursor functions.
• API server cursors are implemented on the server.
• API server cursors support fetching blocks of rows with each fetch.
• A cursor fetches multiple rows at a time is called a block cursor.

Define the steps to use Transact-SQL Cursor.

Declare the cursor,
Open the cursor,
Fetch record row by row,
Close cursor,
Deallocate cursor.

Example of a cursor

Declare @EmpId int
Declare curEmp CURSOR READ_ONLY FOR SELECT EmpId FROM Employee
Open curEmp
Fetch next from curEmp into @EmpId
While @@FETCH_STATUS = 0
Begin Print @EmpId
Fetch next from curEmp into @EmpId
End
Close curEmp
Deallocate curEmp

Explain the cursor types.

DYNAMIC: It reflects changes happened on the table while scrolling through the row.
STATIC: It works on snapshot of record set and disconnects from the server. This kind doesn’t reflects changes happened on the table while scrolling through the row.
KEYSET: In this kind, new record is not reflected, but data modification can be seen.

Define the cursor lock types.

Three types of locks

READ ONLY: This prevents any updates on the table.
SCROLL LOCK: This allows you to make changes to the table.
OPTIMISTIC: This checks if any change in the table record since the row fetched before updating.
If there is no change, the cursor can update.

Explain in brief the cursor optimization tips.

Close cursor when it is not required.
You shouldn’t forget to deallocate cursor after closing it.
You should fetch least number of records.
You should use FORWARD ONLY option when there is no need to update rows.

Explain the disadvantages/limitation of the cursor.

Cursor requires a network roundtrip each time it fetches a record, thus consume network resources.
While data processing, it issues locks on part of the table, or on the whole table.

Define scrollable cursor.

You can use keyword SCROLL to make cursor Scrollable.
It can scroll to any row and can access the same row in the result set multiple times.
A non-scrollable cursor is also known as forward-only and each row can be fetched at most once.

Temporary table VS Table variable: Cursor alternative

Temporary table

This can improve processing speed but consume disk space.

Table variable

Table variable that can be used in stored procedures, functions and batches.
Table variable get destroyed at the end of the stored procedure, function or batch in which it is defined.
Since it can be used in stored procedure, it is compiled once and can be used many times

What does the COLLATE keyword in SQL signify?

Collation is the order that SQL Server uses for sorting or comparing textual data. The temporary database is a system database that inherits it’s collation order from the master. If this temp table is compared to a real one, SQL Server will inform that it cannot resolve collation conflict on = operation.

To solve this, the tables should be created using the COLLATE keyword.

What is COMPUTE and WITH TIES clause in SQL?

The COMPUTE clause is placed at the end of a query to place the result of an aggregate function at the end of a listing.
e.g.: COMPUTE SUM (TOTAL)

The SELECT TOP N query always return exactly N records, and arbitrarily drops any record that have the same value as the last record in the group. The WITH TIES clause can do away with this. It returns the records that have the same value even if the number declared after the TOP has been exceeded