Skip to content

SQL annotations

Jump to bottom
Jean Bisutti edited this page Jul 15, 2019 · 317 revisions

Take control of SQL requests sent to the database


You can take care of several things about SQL requests to favor performance and scalability at the beginning of application development.

Do little configuration described in Quick start before using SQL annotations.

Outline

Quick start

Worflow with SQL annotations

Recommended global annotations

Cancel the behavior of global annotations

Recommended method annotations

Debug annotations

Quick start

Configuration with Spring

Configuration with JUnit 4

Use global annotations or method annotations. See the workflow part to see ways to work with SQL annotations.

The SQL annotations automatically detect if Hibernate or Spring Boot are used. You have no configuration to do.
If a SQL property is unrespected, the SQL annotation can suggest you solutions to fix it with these frameworks.

For example, the following message is diplayed when a N+1 select is presumed and Spring Data JPA is detected:

	* With Spring Data JPA, you may fix it by adding
	@EntityGraph(attributePaths = { "..." }) on repository method.
	https://docs.spring.io/spring-data/jpa/docs/current/reference/html/#jpa.entity-graph

Worflow with SQL annotations

Recommended global annotations

Configure recommended global annotations

A SqlAnnotationBuilder class is available to easily implement SpecifiableAnnotations.

package org.quickperf;

import org.quickperf.config.user.SpecifiableAnnotations;
import org.quickperf.sql.annotation.SqlAnnotationBuilder;

import java.lang.annotation.Annotation;
import java.util.Arrays;
import java.util.Collection;

import static org.quickperf.sql.annotation.SqlAnnotationBuilder.*;

public class QuickPerfConfiguration implements SpecifiableAnnotations {

    public Collection<Annotation> specifyAnnotationsAppliedOnEachTest() {
        return Arrays.asList(  disableSameSelectTypesWithDifferentParams() // can reveal some N+1 selects
                             , disableExactlySameSelects() // can reveal a bad use of Hibernate session
                             , expectJdbcBatching()
                           //, disableCrossJoin() // if you use JPA Criteria API
                             , disableLikeWithLeadingWildcard()
                             );
    }

}

The class implementing SpecifiableAnnotations has to be in org.quickperf package.

@DisableExactlySameSelects

@DisableSameSelectTypesWithDifferentParams

@DisableLikeWithLeadingWildcard

Verify that SQL requests do not contain LIKE with leading wildcard (% or _).

You can read this article explaining why LIKE with leading wildcard could be a bad idea in term of performance.

A code sending to the database a like with leading wilcard may be fast in a test having a few data but very slow with the data volume of production.

@ExpectJdbcBatching

Verify that inserts, deletes and updates are processed in JDBC batches having batchSize elements.

You may sometimes think that you are using JDBC batching but in fact not (Paper 1, Paper 2)!

Batching of inserts, updates and deletes allows to reduce the number of roundtrips to the database which can dramatically impact application performance.

You can decide to batch all inserts, updates, delete. Prior to Hibernate 5.2, batching, when enabled with a hibernate.jdbc.batch_size property stricly positive, was applied to all inserts, updates and deletes (from Hibernate 5.2 it is also possible to override the batch size value for a given session).

Parameters

Parameter Type Meaning Default value
batchSize int JDBC batch size -

batchSize is optional.

A 0 batch size means that JDBC batching is disabled.

Example

    @ExpectJdbcBatching(batchSize = 30)

@DisableCrossJoin

Cross join can be generated by JPA Criteria API and can impact performance.

Cancel the behavior of global annotations

@EnableExactlySameSelects

Cancel behavior of @DisableExactlySameSelects.

Parameters

Parameter Type Meaning Default value
comment String Comment why exactly same selects are enabled -

@EnableSameSelectTypesWithDifferentParams

Cancel behavior of @DisableSameSelectTypesWithDifferentParams.

Parameters

Parameter Type Meaning Default value
comment String Comment why exactly same select types with different parameters are enabled -

@EnableCrossJoin

Cancel behavior of @DisableCrossJoin.

Parameters

Parameter Type Meaning Default value
comment String Comment why cross join is enabled -

@EnableLikeWithLeadingWildcard

Cancel behavior of @DisableLikeWithLeadingWildcard.

Parameters

Parameter Type Meaning Default value
comment String Comment why like with leading wildcard is enabled -

@ExpectJdbcBatching(batchSize=0)

Indicate disabling of JDBC batching.

Recommended method annotations

@ExpectSelect

Parameters

Parameter Type Meaning Default value
value int Number of select requests 0

Example

    @ExpectSelect(1)
    @Test
    public void should_retrieve_all_cars() {	
     //...
    }

@ExpectMaxSelect

With this annotation, the test will fail if the number of SELECT requests is greater than expected.

Parameters

Parameter Type Meaning Default value
value int Maximum number of selects 0

Example

    @ExpectMaxSelect(1)
    @Test
    public void should_retrieve_all_cars() {	
     //...
    }

@ExpectSelectedColumn

Verifies the number of selected columns.

Why limit the number of selected columns?

Parameters

Parameter Type Meaning Default value
value int Number of selected columns 0

Example

    @ExpectSelectedColumn(3)

@ExpectMaxSelectedColumn

With this annotation, the test will fail if the number of returned columns is greater than expected.

Why limit the number of selected columns?

Parameters

Parameter Type Meaning Default value
value int Maximum number of returned columns 0

Example

    @ExpectMaxSelectedColumn(5)

@ExpectInsert

Parameters

Parameter Type Meaning Default value
value int Number of insert requests 0

@ExpectUpdate

Parameters

Parameter Type Meaning Default value
value int Number of update requests 0

@ExpectDelete

Parameter Type Meaning Default value
value int Number of delete requests 0

Debug annotations

@DisplayAppliedAnnotations

@DisplaySqlOfTestMethodBody

With this annotation the SQL orders are diplayed in the console during the execution of the test method body.

Compared to @DisplaySql, this annotation does not diplay SQL orders before (JUnit 4: @Before, @BeforeClass) and after (JUnit 4: @After, @AfterClass) the execution of the test method body.

It is not recommended to commit your test with this annotation. Indeed, the displaying of SQL orders would pollute the logs of your continuous integration build and it may slow down your continuous integration build.

@DisplaySql

With this annotation the SQL orders are diplayed in the console during the test execution.

Compared to @DisplaySqlOfTestMethodBody, this annotation also diplays SQL orders before (JUnit 4: @Before, @BeforeClass) and after (JUnit 4: @After, @AfterClass) the execution of the test method body.

It is not recommended to commit your test with this annotation. Indeed, the displaying of SQL orders would pollute the logs of your continuous integration build and it may slow down your continuous integration build.

:octocat: Code

Twitter URL

📺

QuickPerf

Annotations

👉  Core

👉  JVM

👉  SQL

👉  Scopes

👉  Create an annotation

Supported frameworks

👉  JUnit 4

👉  JUnit 5

👉  TestNG

👉  Spring

How to

👉  Detect and fix N+1 SELECT

Project examples

👉  Maven performance

👉  Spring Boot - JUnit 4

👉  Spring Boot - JUnit 5

👉  Micronaut Data - JUnit 5

👉  Micronaut - Spring - JUnit 5

👉  Quarkus - JUnit 5

Miscellaneous

👉  FAQ

👉  QuickPerf code

Clone this wiki locally