Learn how to use execution plans to analyze and improve your SQL query performance. Find out how to access, read, and fix common issues with scans, joins, sorts, filters, and more.

Sometimes indexes, even though they are present in the table, they aren't used by the SQL engine. Don't be afraid of creating an external table with an indexed column, and join it to your query just to speed things up. Normally, the extra space used by the external table is by far outweighed by the speed your query gains by using this index.

How do you troubleshoot a slow SQL query using execution plans?

If you work with databases, you know how frustrating it can be when a SQL query takes too long to run or returns incorrect results. One of the most effective ways to troubleshoot and optimize your queries is to use execution plans. Execution plans are graphical or textual representations of how the database engine processes your query, showing the steps, costs, and resources involved. In this article, you will learn how to use execution plans to identify and fix common performance issues in your SQL queries.

1 What are execution plans?

Execution plans are generated by the database engine when you run a SQL query, either before or after the execution. They show you the logical and physical operations that the engine performs to retrieve or modify the data, such as scans, joins, sorts, filters, and aggregations. Each operation has an estimated or actual cost, which reflects the amount of CPU, memory, and I/O resources required. Execution plans also show you the data flow between the operations, the number of rows processed, and the indexes used or missing.

Add your perspective

Jameel R.

Inspiring Action | Helping You Crush Your Goals
Report contribution
At MoneyPlus Group, I delved into execution plans to enhance SQL query efficiency. Understanding the steps taken by the database engine—like scans, joins, and sorts—enabled strategic modifications. For instance, if a resource-heavy operation was identified, I'd rework the SQL queries, adjusting joins or optimizing indexes. This hands-on approach significantly improved the efficiency and speed of our database operations, ensuring optimal resource usage while fetching necessary data.

Like

2 How to access execution plans?

Depending on the database management system (DBMS) you use, there are different ways to access execution plans. For example, in Microsoft SQL Server, you can use the graphical execution plan feature in SQL Server Management Studio (SSMS) or the SET STATISTICS XML ON statement to get the XML version of the plan. In Oracle, you can use the EXPLAIN PLAN statement or the DBMS_XPLAN package to get the textual or graphical plan. In MySQL, you can use the EXPLAIN or EXPLAIN EXTENDED statements to get the tabular plan. You should consult the documentation of your DBMS for the specific syntax and options.

Add your perspective

3 How to read execution plans?

When reading execution plans, you should pay attention to the overall cost and duration of the query, the relative cost and percentage of each operation, the number of rows and size of data processed by each operation, the indexes used or missing by each operation, and any warnings or errors displayed by some operations. Depending on the format and preference, execution plans can be read from left to right or from right to left: the rightmost or topmost operation is the first one executed, and the leftmost or bottommost operation is the last one. Each operation has a symbol, a name, and a set of properties that describe its function and cost. Examining these aspects can help you identify potential problems or optimizations and determine how efficient or expensive a query is.

Add your perspective

4 How to troubleshoot slow queries using execution plans?

Using execution plans, you can troubleshoot slow queries by identifying and fixing the root causes of performance issues. Common causes and solutions include table scans or index scans, which can be costly and slow, and can be avoided by using more selective predicates, creating or updating indexes, or rewriting the query. Additionally, nested loops or hash joins can be very CPU-intensive and memory-consuming, so you should reduce the size of the data sets by filtering, indexing, or sorting them, or use a different join type or algorithm. Sorts or aggregations can also be very I/O-intensive and memory-consuming; minimize the amount of data to be sorted or aggregated by filtering, indexing, or grouping them, or use a different sort or aggregate method. Lastly, filters or predicates can be very CPU-intensive and slow; use more efficient and accurate conditions, avoid implicit conversions or functions, or move them to a different place in the query.

Add your perspective

Juan Carlos Munoz Verduzco
Report contribution
Sometimes indexes, even though they are present in the table, they aren't used by the SQL engine. Don't be afraid of creating an external table with an indexed column, and join it to your query just to speed things up. Normally, the extra space used by the external table is by far outweighed by the speed your query gains by using this index.

Like

5 How to optimize queries using execution plans?

Execution plans can optimize queries by applying best practices and techniques to improve performance and readability. It's important to use the appropriate data types and sizes for columns and variables to reduce storage and processing costs and avoid data loss or errors. Consistent naming and formatting conventions for tables, columns, aliases, and queries can enhance readability and maintainability. Comments and documentation should explain the purpose and logic of queries, which can aid understanding and modification. Subqueries, common table expressions (CTEs), views, or stored procedures can simplify and modularize queries, reducing complexity and redundancy. Parameters, variables, or dynamic SQL can make queries more flexible and adaptable to different scenarios, reducing hard-coding and repetition.

Add your perspective

6 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

Add your perspective

Kevin Wood

Experienced leader with experience in implementing technology to achieve organizational goals and overcome challenges
Report contribution
The use of indexes is important (good that you mentioned it). Did you know indexed can become fragmented? After a certain level of fragmentation, the server will no longer use that index you created to speed queries. Defragmenting your indexes and your data are both important (if you want to avoid the slowness you created the index to avoid)

Like

How do you troubleshoot a slow SQL query using execution plans?

1

2

3

4

5

6

1 What are execution plans?

2 How to access execution plans?

3 How to read execution plans?

4 How to troubleshoot slow queries using execution plans?

5 How to optimize queries using execution plans?

6 Here’s what else to consider

System Administration

Rate this article

Thanks for your feedback

More articles on System Administration

More relevant reading

How do you troubleshoot a slow SQL query using execution plans?

1

2

3

4

5

6

1 What are execution plans?

2 How to access execution plans?

3 How to read execution plans?

4 How to troubleshoot slow queries using execution plans?

5 How to optimize queries using execution plans?

6 Here’s what else to consider

System Administration

Rate this article

Thanks for your feedback

Explore Other Skills