Bite Size SQL – Basic Query Optimisation

How SQL queries are written can have a huge impact on how quickly and efficiently they execute. In this post I will explain some very basic methods that can be used to optimise SQL queries.

  • When writing select queries, only select the columns you need. In general using “SELECT *” is very bad practice.
  • Avoid using sub queries, rather use joins. What is meant by this is instead of using:
        SELECT COLUMNA FROM TABLEA WHERE COLUMNB IS IN (SELECT COLUMNB FROM TABLEB)
        Rather use:
        SELECT COLUMNA FROM TABLEA AS TA INNER JOIN TABLEB AS TB ON TA.COLUMNB = TB.COLUMNB
  • Use WHERE to limit the result set to only what you need. There is no point in pulling 10000 records if you are only looking for records, for example in a certain date range.
  • Use WITH NOLOCK with SELECT queries when feasible, this can prevent deadlocks which can cause serious performance issues. However note that using WITH NOLOCK can result in a non-accurate result set being returned which contains non committed records, so keep this in mind when using. But when querying a production system database this risk might be outweighed by the risk of deadlocks occuring.

One last point on stored procedures, this is not really an optimisation point as much as a best practice point. When developing stored procedures always write them to return result sets of the same dimensions. I have worked on numerous legacy systems were this was not the case and it results in a complete nightmare, especially with projects that utilise an ORM.

Bite Size SQL – Basic Query Optimisation

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s