Replace derived `CriteriaQuery` with String-based queries #3653

mp911de · 2024-10-28T07:59:35Z

Seol-JY · 2024-11-07T23:31:39Z

...g-data-jpa/src/main/java/org/springframework/data/jpa/repository/query/ParameterBinding.java

+				switch (type) {
+					case STARTING_WITH:
+						return String.format("%s%%", escape.escape(value.toString()));
+					case ENDING_WITH:
+						return String.format("%%%s", escape.escape(value.toString()));
+					case CONTAINING:
+					case NOT_CONTAINING:
+						return String.format("%%%s%%", escape.escape(value.toString()));
+					default:
+						return value;
+				}


Suggested change

switch (type) {

case STARTING_WITH:

return String.format("%s%%", escape.escape(value.toString()));

case ENDING_WITH:

return String.format("%%%s", escape.escape(value.toString()));

case CONTAINING:

case NOT_CONTAINING:

return String.format("%%%s%%", escape.escape(value.toString()));

default:

return value;

}

String escapedValue = escape.escape(value.toString());

return switch (type) {

case STARTING_WITH -> escapedValue + "%";

case ENDING_WITH -> "%" + escapedValue;

case CONTAINING, NOT_CONTAINING -> "%" + escapedValue + "%";

default -> value;

};

Consider leveraging the switch expression for its readability and declaring escapedValue once to avoid redundancy.

schauder

I left a couple of minor complains and questions

spring-data-jpa/src/main/java/org/springframework/data/jpa/repository/query/JpqlUtils.java

...g-data-jpa/src/main/java/org/springframework/data/jpa/repository/query/ParameterBinding.java

...g-data-jpa/src/main/java/org/springframework/data/jpa/repository/query/PartTreeJpaQuery.java

schauder · 2024-11-20T10:00:52Z

...g-data-jpa/src/main/java/org/springframework/data/jpa/repository/query/PartTreeJpaQuery.java

-					lock.unlock();
+		protected JpqlQueryCreator createCreator(Sort sort, JpaParametersParameterAccessor accessor) {
+
+			synchronized (cache) {


Wouldn't it be less error prone to make the cache implementation synchronized?

It likely does make sense, also the count query requires a revision of its caching to consider nullable values. Good catch!

...a/src/test/java/org/springframework/data/jpa/repository/query/JpqlQueryBuilderUnitTests.java

...org/springframework/data/jpa/repository/query/ParameterMetadataProviderIntegrationTests.java

Introduce new DSL to construct JPQL queries. Refactor ParameterMetadata to PartTreeParameterBinding. Disable Keyset pagination with projections for Eclipselink as Eclipselink doesn't consider type hints for JPQL queries.

Make usage of ParameterExpression more explicit. Add JPQL rendering tests. Favor Metamodel over From for building jpql queries. Align IsNull and IsNotNull handling. Support Derived Delete and Exists, consider null values when caching queries.

Remove method overloads accepting pure strings. Use switch-expressions. Correctly navigate nested joins. Introduce PathExpression interface, refine naming.

Introduce new DSL to construct JPQL queries. Refactor ParameterMetadata to PartTreeParameterBinding. Disable Keyset pagination with projections for Eclipselink as Eclipselink doesn't consider type hints for JPQL queries. Closes #3588 Original pull request: #3653