912 : SchemaMigrator/SchemaValidator support for MS SQL Server

README.md

@@ -36,10 +36,10 @@ Hibernate Reactive has been tested with:
 - CockroachDB 21.1
 - MS SQL Server 2019
 - [Hibernate ORM][] 5.6.0.Beta2
-- [Vert.x Reactive PostgreSQL Client](https://vertx.io/docs/vertx-pg-client/java/) 4.1.3
-- [Vert.x Reactive MySQL Client](https://vertx.io/docs/vertx-mysql-client/java/) 4.1.3
-- [Vert.x Reactive Db2 Client](https://vertx.io/docs/vertx-db2-client/java/) 4.1.3
-- [Vert.x Reactive MS SQL Server Client](https://vertx.io/docs/vertx-mssql-client/java/) 4.1.3
+- [Vert.x Reactive PostgreSQL Client](https://vertx.io/docs/vertx-pg-client/java/) 4.1.4
+- [Vert.x Reactive MySQL Client](https://vertx.io/docs/vertx-mysql-client/java/) 4.1.4
+- [Vert.x Reactive Db2 Client](https://vertx.io/docs/vertx-db2-client/java/) 4.1.4
+- [Vert.x Reactive MS SQL Server Client](https://vertx.io/docs/vertx-mssql-client/java/) 4.1.4
 - [Quarkus][Quarkus] via the Hibernate Reactive extension
 
 [PostgreSQL]: https://www.postgresql.org

build.gradle

@@ -79,7 +79,7 @@ ext {
     // Example:
     // ./gradlew build -PvertxVersion=4.0.0-SNAPSHOT
     if ( !project.hasProperty('vertxVersion') ) {
-        vertxVersion = '4.1.3'
+        vertxVersion = '4.1.4'
     }
 
     testcontainersVersion = '1.16.0'

gradle.properties

@@ -29,21 +29,21 @@ org.gradle.java.installations.auto-download=false
 #db = MSSQL
 
 # Enable the SonatypeOS maven repository (mainly for Vert.x snapshots) when present (value ignored)
-#enableSonatypeOpenSourceSnapshotsRep = true
+enableSonatypeOpenSourceSnapshotsRep = true
 
 # Enable the JBoss Snapshot maven repository (mainly for Hibernate ORM snapshots) when present (value ignored)
 #enableJBossSnapshotsRep = true
 
 # Enable the maven local repository (for local development when needed) when present (value ignored)
-#enableMavenLocalRepo = true
+enableMavenLocalRepo = true
 
 # Override default Hibernate ORM version
-#hibernateOrmVersion = 5.5.6-SNAPSHOT
+hibernateOrmVersion = 5.6.0-SNAPSHOT
 
 # If set to true, skip Hibernate ORM version parsing (default is true, if set to null)
 # this is required when using intervals or weird versions or the build will fail
 #skipOrmVersionParsing = true
 
 # Override default Vert.x version
-#vertxVersion = 4.1.3-SNAPSHOT
+#vertxVersion = 4.1.4-SNAPSHOT
 

hibernate-reactive-core/src/main/java/org/hibernate/reactive/provider/service/AbstractReactiveInformationSchemaBasedExtractorImpl.java

@@ -130,6 +130,11 @@ private String normalizeLabel(String columnLabel) {
 		return toMetaDataObjectName( Identifier.toIdentifier( columnLabel ) );
 	}
 
+	@Override
+	protected String getResultSetTableTypesPhysicalTableConstant() {
+		return "BASE TABLE";
+	}
+
 	@Override
 	protected <T> T processCatalogsResultSet(ExtractionContext.ResultSetProcessor<T> processor) throws SQLException {
 		return getExtractionContext().getQueryResults(
@@ -150,7 +155,7 @@ protected <T> T processSchemaResultSet(
 		final StringBuilder sb = new StringBuilder()
 				.append( "select catalog_name as " ).append( getResultSetCatalogLabel() )
 				.append( " , schema_name as "  ).append( getResultSetSchemaLabel() )
-				.append( " from information_schema.schemata where true" );
+				.append( " from information_schema.schemata where 1 = 1" );
 		final List<Object> parameters = new ArrayList<>();
 		appendClauseAndParameterIfNotNullOrEmpty( " and catalog_name = ", catalog, sb, parameters );
 		appendClauseAndParameterIfNotNullOrEmpty( " and schema_name like ", schemaPattern, sb, parameters );
@@ -198,7 +203,7 @@ protected <T> T processTableResultSet(
 				// Remarks are not available from information_schema.
 				// Hibernate ORM does not currently do anything with remarks,
 				// so just return null for now.
-				.append( " from information_schema.tables where true" );
+				.append( " from information_schema.tables where 1 = 1" );
 		List<Object> parameterValues = new ArrayList<>();
 		appendClauseAndParameterIfNotNullOrEmpty( " and " + catalogColumn + " = ", catalog, sb, parameterValues );
 		appendClauseAndParameterIfNotNullOrEmpty( " and " + schemaColumn + " like ", schemaPattern, sb, parameterValues );
@@ -255,7 +260,7 @@ protected <T> T processColumnsResultSet(
 				// Hibernate's metadata for the column. ORM also considers
 				// the same column type name as a match, so the SQL code is
 				// optional. For now, just return null for the SQL type code.
-				.append( " from information_schema.columns where true" );
+				.append( " from information_schema.columns where 1 = 1" );
 
 		final List<Object> parameterValues = new ArrayList<>();
 		final String catalogColumn = getDatabaseCatalogColumnName(

hibernate-reactive-core/src/main/java/org/hibernate/reactive/provider/service/ReactiveSchemaManagementTool.java

@@ -12,6 +12,7 @@
 import org.hibernate.dialect.MariaDB103Dialect;
 import org.hibernate.dialect.MySQL8Dialect;
 import org.hibernate.dialect.PostgreSQL10Dialect;
+import org.hibernate.dialect.SQLServer2012Dialect;
 import org.hibernate.engine.jdbc.env.spi.JdbcEnvironment;
 import org.hibernate.resource.transaction.spi.DdlTransactionIsolator;
 import org.hibernate.service.ServiceRegistry;
@@ -67,6 +68,9 @@ public InformationExtractor createInformationExtractor(ExtractionContext extract
 			else if ( dialect instanceof MySQL8Dialect || dialect instanceof MariaDB103Dialect ) {
 				return new MySqlReactiveInformationExtractorImpl( extractionContext );
 			}
+			else if ( dialect instanceof SQLServer2012Dialect ) {
+				return new SqlServerReactiveInformationExtractorImpl( extractionContext );
+			}
 			else {
 				throw new NotYetImplementedException(
 						"No InformationExtractor for Dialect [" + dialect + "] is implemented yet"

hibernate-reactive-core/src/main/java/org/hibernate/reactive/provider/service/SqlServerReactiveInformationExtractorImpl.java

@@ -0,0 +1,252 @@
+/* Hibernate, Relational Persistence for Idiomatic Java
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ * Copyright: Red Hat Inc. and Hibernate Authors
+ */
+package org.hibernate.reactive.provider.service;
+
+import java.sql.SQLException;
+import java.util.ArrayList;
+import java.util.List;
+
+import org.hibernate.boot.model.naming.Identifier;
+import org.hibernate.cfg.NotYetImplementedException;
+import org.hibernate.tool.schema.extract.spi.ExtractionContext;
+
+/**
+ * An implementation of {@link AbstractReactiveInformationSchemaBasedExtractorImpl}
+ * specifically for SQL Server that obtains metadata from SQL Server's system
+ * tables, when it is not available from SQL Server's information_schema.
+ *
+ * @author Gail Badner
+ */
+public class SqlServerReactiveInformationExtractorImpl extends AbstractReactiveInformationSchemaBasedExtractorImpl  {
+
+	public SqlServerReactiveInformationExtractorImpl(ExtractionContext extractionContext) {
+		super( extractionContext );
+	}
+
+	// #processSchemaResultSet in the superclass is OK as is
+	// because it uses information_schema.schemata and it is the
+	// schema itself that is the data object. Documentation says
+	// the schema owner may incomplete, but the schema owner is not
+	// needed.
+	// (See https://docs.microsoft.com/en-us/sql/relational-databases/system-information-schema-views/schemata-transact-sql?view=sql-server-ver15&viewFallbackFrom=sql-server-ver19)
+
+	@Override
+	protected <T> T processTableResultSet(
+			String catalog,
+			String schemaPattern,
+			String tableNamePattern,
+			String[] types,
+			ExtractionContext.ResultSetProcessor<T> processor
+	) throws SQLException {
+
+		// Documentation for information_schema.tables says the following for
+		// table_schema:
+		// "** Important ** only reliable way to find the schema of an object
+		// is to query the sys.objects catalog view. INFORMATION_SCHEMA views
+		// could be incomplete since they are not updated for all new features."
+		// (See https://docs.microsoft.com/en-us/sql/relational-databases/system-information-schema-views/tables-transact-sql?view=sql-server-ver15&viewFallbackFrom=sql-server-ver19)
+
+		// This is the reason for joining information_schema.tables with sys.objects.
+
+		final StringBuilder sb = new StringBuilder()
+				.append( "select t.table_catalog as " ).append( getResultSetCatalogLabel() )
+				.append( " , OBJECT_SCHEMA_NAME( o.object_id ) as " ).append( getResultSetSchemaLabel() )
+				.append( " , t.table_name as " ).append( getResultSetTableNameLabel() )
+				.append( " , t.table_type as " ).append( getResultSetTableTypeLabel() )
+				.append( " , null as " ).append( getResultSetRemarksLabel() )
+				// Remarks are not available from information_schema.
+				// Hibernate ORM does not currently do anything with remarks,
+				// so just return null for now.
+				.append( " from information_schema.tables t inner join sys.objects o" )
+				.append( " on t.table_name = o.name " )
+				.append( " and ( ( t.table_type = 'BASE TABLE' and o.type = 'U' ) or ( t.table_type = 'VIEW' and o.type = 'V' ) )" )
+				// o.type = 'U' is for a user-defined table
+				// o.type = 'V' is for a view
+				.append( " where 1 = 1" );
+
+		List<Object> parameterValues = new ArrayList<>();
+
+		appendClauseAndParameterIfNotNullOrEmpty( " and t.table_catalog = ", catalog, sb, parameterValues );
+		appendClauseAndParameterIfNotNullOrEmpty( " and OBJECT_SCHEMA_NAME( o.object_id ) like ", schemaPattern, sb, parameterValues );
+		appendClauseAndParameterIfNotNullOrEmpty( " and t.table_name like ", tableNamePattern, sb, parameterValues );
+
+		if ( types != null && types.length > 0 ) {
+			appendClauseAndParameterIfNotNullOrEmpty(
+					" and t.table_type in ( ",
+					types[0].equals( "TABLE" ) ? getResultSetTableTypesPhysicalTableConstant() : types[0],
+					sb,
+					parameterValues
+			);
+			for ( int i = 1 ; i < types.length ; i++ ) {
+				appendClauseAndParameterIfNotNullOrEmpty(
+						", ",
+						types[i].equals( "TABLE" ) ? getResultSetTableTypesPhysicalTableConstant() : types[i],
+						sb,
+						parameterValues
+				);
+			}
+			sb.append( " ) " );
+		}
+		return getExtractionContext().getQueryResults( sb.toString(), parameterValues.toArray(), processor );
+	}
+
+	@Override
+	protected <T> T processColumnsResultSet(
+			String catalog,
+			String schemaPattern,
+			String tableNamePattern,
+			String columnNamePattern,
+			ExtractionContext.ResultSetProcessor<T> processor) throws SQLException {
+
+		// Documentation for information_schema.columns says the following for
+		// table_schema:
+		// ** Important ** Do not use INFORMATION_SCHEMA views to determine the
+		// schema of an object. INFORMATION_SCHEMA views only represent a subset
+		// of the metadata of an object. The only reliable way to find the schema
+		// of a object is to query the sys.objects catalog view.
+		// (See https://docs.microsoft.com/en-us/sql/relational-databases/system-information-schema-views/columns-transact-sql?view=sql-server-ver15&viewFallbackFrom=sql-server-ver19)
+
+		final StringBuilder sb = new StringBuilder()
+				.append( "select c.table_name as " ).append( getResultSetTableNameLabel() )
+				.append( ", c.column_name as " ).append( getResultSetColumnNameLabel() )
+				.append( ", c." ).append( getInformationSchemaColumnsDataTypeColumn() )
+				.append( " as " ).append( getResultSetTypeNameLabel() )
+				.append( ", null as " ).append( getResultSetColumnSizeLabel() )
+				// Column size is fairly complicated to get out of information_schema
+				// and likely to be DB-dependent. Currently, Hibernate ORM does not use
+				// column size for anything, so for now, just return null.
+				.append( ", null as " ) .append( getResultSetDecimalDigitsLabel() )
+				// Decimal digits is fairly complicated to get out of information_schema
+				// and likely to be DB-dependent. Currently, Hibernate ORM does not use
+				// decimal digits for anything, so for now, just return null.
+				.append( ", c.is_nullable as " ).append( getResultSetIsNullableLabel() )
+				.append( ", null as " ).append( getResultSetSqlTypeCodeLabel() )
+				// There is a SQL type code available from sys.types in SQL Server,
+				// Currently, Hibernate ORM only uses
+				// the SQL type code for SchemaMigrator to check if a column
+				// type in the DB is consistent with what is computed in
+				// Hibernate's metadata for the column. ORM also considers
+				// the same column type name as a match, so the SQL code is
+				// optional. For now, just return null for the SQL type code.
+				.append( " from information_schema.columns c inner join sys.objects o" )
+				.append( " on c.table_name = o.name and o.type in ( 'U', 'V' ) " )
+				// o.type = 'U' is for a user-defined table
+				// o.type = 'V' is for a view
+				.append( " where 1 = 1" );
+
+		final List<Object> parameterValues = new ArrayList<>();
+		appendClauseAndParameterIfNotNullOrEmpty( " and c.table_catalog = " , catalog, sb, parameterValues );
+		appendClauseAndParameterIfNotNullOrEmpty( " and OBJECT_SCHEMA_NAME( o.object_id ) like " , schemaPattern, sb, parameterValues );
+		appendClauseAndParameterIfNotNullOrEmpty( " and c.table_name like " , tableNamePattern, sb, parameterValues );
+
+		sb.append(  " order by c.table_catalog, OBJECT_SCHEMA_NAME( o.object_id ), c.table_name, c.column_name, c.ordinal_position" );
+
+		return getExtractionContext().getQueryResults( sb.toString(), parameterValues.toArray(), processor );
+	}
+
+	@Override
+	protected <T> T processPrimaryKeysResultSet(
+			String catalogFilter,
+			String schemaFilter,
+			Identifier tableName,
+			ExtractionContext.ResultSetProcessor<T> processor) throws SQLException {
+		throw new NotYetImplementedException();
+	}
+
+	@Override
+	protected <T> T processIndexInfoResultSet(
+			String catalog,
+			String schema,
+			String table,
+			boolean unique,
+			boolean approximate,
+			ExtractionContext.ResultSetProcessor<T> processor) throws SQLException {
+
+		StringBuilder sb = new StringBuilder()
+				.append( "select i.name as " ).append( getResultSetIndexNameLabel() )
+				.append( " , i.type as " ).append( getResultSetIndexTypeLabel() )
+				.append( " , COL_NAME(ic.object_id, ic.column_id) as " ).append( getResultSetColumnNameLabel() )
+				.append( " from sys.indexes i inner join sys.index_columns ic" )
+				.append( " on ic.object_id = i.object_id and ic.index_id = i.index_id" )
+				.append( " where i.index_id > 0" )
+				.append( " and i.type in (1, 2)" )
+				.append( " and i.is_primary_key = 0" );
+				// do not include PK indexes
+
+		final List<Object> parameterValues = new ArrayList<>();
+
+		appendClauseAndParameterIfNotNullOrEmpty( " and DB_NAME() = ", catalog, sb, parameterValues );
+		appendClauseAndParameterIfNotNullOrEmpty( " and OBJECT_NAME( i.object_id ) = ", table, sb, parameterValues );
+		appendClauseAndParameterIfNotNullOrEmpty( " and OBJECT_SCHEMA_NAME( i.object_id) = ", schema, sb, parameterValues );
+
+		if ( unique ) {
+			sb.append( " and i.is_unique_constraint = true" );
+		}
+
+		sb.append( " order by OBJECT_SCHEMA_NAME( i.object_id), OBJECT_NAME( i.object_id ), ic.key_ordinal" );
+
+		T result = getExtractionContext().getQueryResults(
+				sb.toString(),
+				parameterValues.toArray(),
+				processor
+		);
+		return result;
+	}
+
+	@Override
+	protected <T> T processImportedKeysResultSet(
+			String catalog,
+			String schema,
+			String table,
+			ExtractionContext.ResultSetProcessor<T> processor) throws SQLException {
+
+		// Documentation for information_schema.key_column_usage says the following for
+		// table_schema and constraint_schema:
+
+		// ** Important ** Do not use INFORMATION_SCHEMA views to determine the schema
+		// of an object. INFORMATION_SCHEMA views only represent a subset of the
+		// metadata of an object. The only reliable way to find the schema of a object
+		// is to query the sys.objects catalog view.
+		// (See https://docs.microsoft.com/en-us/sql/relational-databases/system-information-schema-views/key-column-usage-transact-sql?view=sql-server-ver15&viewFallbackFrom=sql-server-ver19)
+
+		// The ResultSet must be ordered by the primary key catalog/schema/table and column position within the key.
+
+		final StringBuilder sb = new StringBuilder()
+				.append( "select OBJECT_NAME( constraint_object_id ) as " ).append( getResultSetForeignKeyLabel() )
+				.append( ", DB_NAME() as " ).append( getResultSetPrimaryKeyCatalogLabel() )
+				.append( ", OBJECT_SCHEMA_NAME( referenced_object_id ) as " ).append( getResultSetPrimaryKeySchemaLabel() )
+				.append( ", OBJECT_NAME( referenced_object_id ) as " ).append( getResultSetPrimaryKeyTableLabel() )
+				.append( ", COL_NAME( parent_object_id, parent_column_id ) as ").append( getResultSetForeignKeyColumnNameLabel() )
+				.append( ", COL_NAME( referenced_object_id, referenced_column_id) as ").append( getResultSetPrimaryKeyColumnNameLabel() )
+				.append( " from sys.foreign_key_columns" )
+				.append( " where 1 = 1" );
+
+		// Now add constraints for the requested catalog/schema/table
+
+		final List<Object> parameters = new ArrayList<>();
+		final List<String> orderByList = new ArrayList<>();
+
+		if ( appendClauseAndParameterIfNotNullOrEmpty( " and DB_NAME() = ", catalog, sb, parameters ) ) {
+			orderByList.add( "DB_NAME()" );
+		}
+		if ( appendClauseAndParameterIfNotNullOrEmpty( " and OBJECT_SCHEMA_NAME( parent_object_id ) = ", schema, sb, parameters ) ) {
+			orderByList.add( "OBJECT_SCHEMA_NAME( parent_object_id )" );
+		}
+		if ( appendClauseAndParameterIfNotNullOrEmpty( " and OBJECT_NAME( parent_object_id ) = ", table, sb, parameters ) ) {
+			orderByList.add( "OBJECT_NAME( parent_object_id )" );
+		}
+		orderByList.add( "constraint_column_id" );
+
+		if ( orderByList.size() > 0 ) {
+			sb.append( " order by " ).append( orderByList.get( 0 ) );
+			for ( int i = 1 ; i < orderByList.size() ; i++ ) {
+				sb.append( ", " ).append( orderByList.get( i ) );
+			}
+		}
+
+		return getExtractionContext().getQueryResults( sb.toString(), parameters.toArray(), processor );
+	}
+}