Develop Reference

Difference between spring.jpa.hibernate.hbm2ddl and spring.jpa.hibernate.ddl

What is the difference between spring.jpa.hibernate.hbm2ddl and spring.jpa.hibernate.ddl?
I have found in this question: What are the possible values of spring.datasource.initialization-mode? that OP is using both in properties, however it seems like the origin of hbm2ddl is hibernate directly not Spring Data Jpa.
Nevertheless, reading the answer from another OP, it looks like pass-through only.
However in our commercial project with mariadb, when we do not close our spring boot application gracefully with spring.jpa.hibernate.hbm2ddl.auto=create, when the application is run again, it deletes old data and creates everything from scratch. On the other hand with spring.jpa.hibernate.ddl.auto=create every second run (after no graceful application shutdown) causes key constraint exceptions (DB is not being dropper before creation)

From this Link
By default, JPA databases are automatically created only if you use an embedded database (H2, HSQL, or Derby).
You can explicitly configure JPA settings by using spring.jpa.* properties. For example, to create and drop tables you can add the following line to your application.properties:
spring.jpa.hibernate.ddl-auto=create-drop
Hibernate’s own internal property name for this (if you happen to remember it better) is hibernate.hbm2ddl.auto.
From this Link
spring.jpa.hibernate.ddl-auto This is actually a shortcut for the "hibernate.hbm2ddl.auto" property.
Defaults to "create-drop" when using an embedded database and no schema manager was detected. Otherwise, defaults to "none".
From this Link
Spring Boot can automatically create the schema (DDL scripts) of your DataSource and initialize it (DML scripts).
It loads SQL from the standard root classpath locations: schema.sql and data.sql, respectively.
In addition, Spring Boot processes the schema-${platform}.sql and data-${platform}.sql files (if present), where platform is the value of spring.datasource.platform.
This allows you to switch to database-specific scripts if necessary. For example, you might choose to set it to the vendor name of the database (hsqldb, h2, oracle, mysql, postgresql, and so on).

Make schema created by Spring JDBC the default schema

I'm using Spring Boot and JDBC for my database connection. I placed schema.sql at the classpath to initialize a schema and tables.
Because the schema doesn't exist yet while connecting to the datasource, I have to configure the datasource in application.properties like so:
spring.datasource.url=jdbc:mysql://localhost:3306/
schema.sql:
CREATE DATABASE IF NOT EXISTS <schema_name>
USE <schema.name>;
CREATE TABLE...
So I select the schema after creating it. This obviously doesn't persist for too long.
How do I configure this properly? Is there a way to select a default schema after the create script or maybe change the datasource url?

With JDBC you need to use Connection.setCatalog to switch between databases. You should not use USE <databasename> as the JDBC driver itself needs to be aware of which database it is operating on.

Based on your code from the schema.sql
USE <schema.name>;
This will not work for your java environment. The schema.sql will be executed and finished, which will not cater your requirement to set the default schema.
The General approach will be to use JDBC URL as;
jdbc:mysql://localhost:3306/DB_NAME
This will set the default db as DB_NAME.
Assumptions: I am assuming that you want to connect to single node DB without any loadbalancer or failover mechanism to be used. URL may change based on these features to be configured.
If you are not specifying the DB_NAME in the URL, it means their is no default schema.
You have 2 options to access the DB in that case.
1) Always use the Connection.setCatalog() method to specify the desired database in JDBC applications, rather than the USE database statement.
2) fully specify table names using the database name (that is, SELECT dbname.tablename.colname FROM dbname.tablename...) in your SQL. Opening a connection without specifying the database to use is generally only useful when building tools that work with multiple databases, such as GUI database managers.
For more details refer to the below mysql portal for reference.
https://dev.mysql.com/doc/connector-j/en/connector-j-reference-configuration-properties.html

mysql dialect in hsqldb

I'm using hibernate to manage DB operations and MySQL in my application - use org.hibernate.dialect.MySQLInnoDBDialect as the hibernate dialect. But for testing purposes I'm using HSQLDB 1.8.0.10.
I have problem with query like this (working good on mysql not on hsql):
SELECT DISTINCT(id) FROM table ORDER BY name;
I know that the problem is with distinct and order by (http://weblogs.sqlteam.com/jeffs/archive/2007/12/13/select-distinct-order-by-error.aspx ) and solution of this could be for example:
SELECT DISTINCT(id) FROM table GROUP BY id ORDER BY MAX(name);
But my question is, if there is any possibility to using MySQL dialect in HSQLDB and not have to using this solution?

HSQLDB implements the SQL Standard correctly and does not allow the ambiguous query. It is not possible to change its behaviour.
It is better to modify your MySQL queries to be standard compliant. This allows you to port your application to another database more easily.

Does Spring embedded database support different SQL dialects?

H2 has a range of compatibility modes for various other databases such as MS SQL Server, MySQL, Oracle, etc that support different SQL dialects. However, when setting up an embedded database in Spring I do not find any corresponding setting. Does this imply that I have to use "plain" SQL without any dialect specific features if I for example use Oracle in production and H2 during test? Have I overlooked something?

which version of H2 database? per the documentation, you can set compatible mode by SQL statement (http://www.h2database.com/html/features.html#compatibility)
SET MODE PostgreSQL
just add this statement into your first sql script file loaded by Spring jdbc embedded-database

According to the H2 doc, the Oracle compatibility mode is quite limited.
For instance, you can not use PL/SQL procedures.
If you use Spring's EmbeddedDatabase, you cannot set the compatibility mode as-is; you have to implement you own EmbeddedDatabaseConfigurer and specify the compatibility mode through the JDBC URL (see below).
But also, to use the compatibility mode with H2 and Spring, you just have to set the mode in your JDBC URL (so it is not Spring related) in a classic way, using a DataSource:
jdbc:h2:~/test;MODE=Oracle
And if you use Hibernate, you have to specify the Oracle dialect instead of the H2 one.

You have 2 options:
use spring to start the H2 database as follows (check setName() to see how to pass H2 specific URL parameters to spring builder):
Spring code generates the URL as follows:
String.format("jdbc:h2:mem:%s;DB_CLOSE_DELAY=-1", databaseName)
So, in setName() you can all any H2 specific parameter in the URL.
private DataSource dataSource() {
EmbeddedDatabaseBuilder builder = new EmbeddedDatabaseBuilder();
EmbeddedDatabase db = builder
.setType(EmbeddedDatabaseType.H2)
.setName("testdb;DATABASE_TO_UPPER=false;MODE=Oracle")
.addScript("schema.sql")
.addScript("data.sql")
.build();
return db;
}
configure directly the DB URL, sth like:
org.h2.jdbcx.JdbcDataSource dataSource = new org.h2.jdbcx.JdbcDataSource();
dataSource.setURL("jdbc:h2:testdb;MODE=MySQL;DATABASE_TO_UPPER=false;INIT=runscript from 'src/test/resources/schema.sql'\;runscript from 'src/test/resources/data.sql'");
The main different is that (2) is executing scripts defined at INIT for every database connection creation and not once per DB creation! This causes various issues, like INSERTs failing due to duplicate keys etc..

Are there any in-memory JDBC databases that support XA distributed transactions?

I want to use an in-memory database to test my application, but it needs to support XA distributed transactions. My first idea for an in-memory database was HSQLDB, but it appears not to support XA. Are there any?

Looks like H2 supports this.

Both H2 database and HSQLDB support it, couldn't find any other java embedded databases even in 2019. But it's not correctly implemented in either of them: the transaction is rolled back as soon as the client disconnects. According to the X/Open spec, when a database prepares a transaction, its data should be persistently stored and must be available to be committed later.
In H2 the XA code is unmaintained.
If you want an XA-supporting database for your tests, you can use Postgres using the TestContainers project:
#ClassRule
public static PostgreSQLContainer container = new PostgreSQLContainer<>("postgres:12.1")
.withCommand("postgres -c max_prepared_transactions=10");
Then create a datasource like this:
PGXADataSource dataSource = new PGXADataSource();
dataSource.setURL(container.getJdbcUrl());
dataSource.setUser(container.getUsername());
dataSource.setPassword(container.getPassword());
dataSource.setDatabaseName(container.getDatabaseName());