SQL Server Statistics

Why should we be interested in SQL Server Statistics, when we just want to get data from a database? The answer is ‘performance’. The better the information that SQL Server has about the data in a database, the better its choices will be on how it executes the SQL. Statistics are its’ chief source of information. If this information is out of date, performance of queries will suffer. The SQL queries that you execute are first passed to the SQL Server Query Optimizer. If it doesn’t have a plan already stored or ‘cached’, it will create an execution plan. When it is creating an execution plan, the SQL Server optimizer chooses an appropriate physical operator for any logical operation (such as a join, or a search) to perform. When doing so, the optimizer has some choice in the way that it maps a distinct physical operator to a logical operation. The optimizer can, for example, select either a ‘physical’ Index Seek or Table Scan when doing a ‘logical’ search, or could choose between a Hash, - Merge-, and Nested Loop joins in order to do an inner join.