Faced with SQL query performance problems that contain a large number of OR conditions, the answer is to optimize by reducing the number of ORs, using indexes reasonably, and adjusting the structure. Specific methods include: 1. Split the query into multiple subqueries and merge them with UNION or UNION ALL to enable each subquery to use indexes independently; 2. Use IN to replace multiple OR conditions in the same field to improve readability and execution efficiency; 3. Create appropriate indexes, such as single column index, composite index or overlay index to accelerate data retrieval; 4. Optimize from the data modeling level, such as introducing a tag system, intermediate table or replacing OR conditions with JOIN, thereby fundamentally reducing the use of OR.

Performance tends to get very poor when you face a SQL query with a large number of OR conditions. These conditions usually cause the database to abandon the use of indexes, resulting in full table scanning, which in turn affects the response speed. The key to optimizing such queries is to reduce the number of ORs, use indexes reasonably to speed up retrieval, or make structural adjustments.

Here are some practical optimization methods:

1. Split the query into multiple subqueries and merge with UNION

When there are many conditions for OR joins in a query, especially when different fields or different parts of a table, consider splitting them into multiple independent SELECT statements and splicing them with UNION or UNION ALL .

Why does this work?
Each subquery can use the index separately, without having to let the optimizer abandon the use of indexes because of too many OR conditions.

For example:

 SELECT * FROM orders WHERE customer_id = 100 OR customer_id = 200 OR customer_id = 300;

It can be rewritten as:

 SELECT * FROM orders WHERE customer_id = 100
UNION ALL
SELECT * FROM orders WHERE customer_id = 200
UNION ALL
SELECT * FROM orders WHERE customer_id = 300;

Note: If the results may be repeated, use UNION ; if you are sure that there will be no duplication, use UNION ALL faster.

2. Use IN to replace multiple OR conditions

If your OR is used to compare multiple values ??in the same field, the most direct way is to use IN instead.

for example:

 SELECT * FROM users WHERE id = 1 OR id = 2 OR id = 3 OR id = 4;

It can be rewritten as:

 SELECT * FROM users WHERE id IN (1, 2, 3, 4);

This approach is not only simpler, but most databases are more efficient in processing IN , especially when there are indexes.

3. Create a suitable composite or overlay index

Sometimes even if you use IN or split the query, if the relevant fields do not have a suitable index, the performance is still not ideal.

You can try to create the following indexes based on the fields in the query:

• Single column index (for IN or equivalent query for a single field)
• Composite index (suitable for multiple fields joint query)
• Overwrite index (including all fields required for the query)

For example, if you often execute queries like this:

 SELECT name FROM users WHERE status = 'active' OR status = 'pending';

Then try to create an overlay index:

 CREATE INDEX idx_users_status_name ON users(status, name);

This way the database can get data directly from the index without having to return to the table.

4. Consider optimization at the data modeling level

If a query requires dozens or even hundreds of OR conditions, this may itself indicate that there is something wrong with your data model design.

for example:

• Should data for certain OR conditions be classified as labeling systems?
• Is it possible to replace multiple ORs with intermediate tables?
• Is it possible to convert certain conditions to JOIN?

For example, suppose you have an order table and often need to query orders by multiple user roles:

 SELECT * FROM orders WHERE user_role = 'admin' OR user_role = 'editor' OR user_role = 'manager';

A better approach at this time might be to introduce a "permission group" table, using JOIN instead of the OR list.

Basically these common optimization methods. The scenarios applicable to each method are slightly different. Which one to choose depends on the actual query structure and data distribution. But the overall idea is: reduce the number of ORs, make good use of indexes, and reconstruct query logic if necessary.

The above is the detailed content of How to optimize a query with too many OR conditions?. For more information, please follow other related articles on the PHP Chinese website!