How to update MySQL database to keep up with Hibernate 6? Previously with Hibernate 5 I had only one sequence named hibernate_sequence that was used to generate IDs.
On my test server Hibernate 6 created sequence for each table storing my entities. It was easy because there were no data and Hibernate was allowed to drop, create, do whatever it needs.
Next I tried to manually upgrade the copy of production database by creating sequences:
DROP TABLE hibernate_sequence;
CREATE TABLE article_seq ( next_val BIGINT(20) );
INSERT INTO article_seq ( next_val ) VALUES ( 1 );
UPDATE article_seq SET next_val = (
SELECT MAX(id) + 1
FROM article
);
-- and so on for each table
The app is working fine in read only mode (not really read only mode but whe used to display existing articles in database) but first time I try to persist new item, I get an error about duplicate keys.
The idea is to keep up with new Hibernate versions and not to use Hibernate 6 strategy legacy which would fix my problem.
What should I do to keep my production data and have new sequence naming conventions applied?
EDIT:
Conversation moved to Hibernate forum:
https://discourse.hibernate.org/t/hibernate-6-sequece-maybe-not-working-as-expected/7228/3
Related
I am using JPA to generate a script for creating database tables based on my entities:
javax.persistence.schema-generation.scripts.action=create
javax.persistence.schema-generation.scripts.create-target=db_setup.sql
The file is generated with the correct tables, however the statements do not end with a semicolon, for example:
create table hibernate_sequence (next_val bigint)
insert into hibernate_sequence values ( 1 )
insert into hibernate_sequence values ( 1 )
I assume this is valid in standard SQL? However, it is not valid in MySQL. Is there a way to tell JPA to add the semicolons to the end of the line? Or what else could be the reason that it is missing?
Since Hibernate 5.1.0, the line delimiter for the generated SQL can be defined by setting the hibernate.hbm2ddl.delimiter property, e.g.
hibernate.hbm2ddl.delimiter=";"
The default is no delimiter.
See also comments on this ticket.
I have an update query which I am trying to execute through batchUpdate method of spring jdbc template. This update query can potentially match 1000s of rows in EVENT_DYNAMIC_ATTRIBUTE table which needs to be get updated. Will updating thousands of rows in a table cause any issue in production database apart from timeout? like, will it crash database or slowdown the performance of entire database engine for other connections...etc?
Is there a better way to achieve this instead of firing single update query in spring JDBC template or JPA? I have the following settings for jdbc template.
this.jdbc = new JdbcTemplate(ds);
jdbc.setFetchSize(1000);
jdbc.setQueryTimeout(0); // zero means there is no limit
The update query:
UPDATE EVENT_DYNAMIC_ATTRIBUTE eda
SET eda.ATTRIBUTE_VALUE = 'claim',
eda.LAST_UPDATED_DATE = SYSDATE,
eda.LAST_UPDATED_BY = 'superUsers'
WHERE eda.DYNAMIC_ATTRIBUTE_NAME_ID = 4002
AND eda.EVENT_ID IN
(WITH category_data
AS ( SELECT c.CATEGORY_ID
FROM CATEGORY c
START WITH CATEGORY_ID = 495984
CONNECT BY PARENT_ID = PRIOR CATEGORY_ID)
SELECT event_id
FROM event e
WHERE EXISTS
(SELECT 't'
FROM category_data cd
WHERE cd.CATEGORY_ID = e.PRIMARY_CATEGORY_ID))
If it is one time thing, I normally first select the records which needs to be updated and put in a temporary table or in a csv, and I make sure that I save primary key of those records in a table or in a csv. Then I read records in batches from temporary table or csv, and do the update in the table using the primary key. This way tables are not locked for a long time and you can have fixed set of records added in the batch which needs update and updates are done using primary key so it will be very fast. And if any update fails then you know which records got failed by logging out the failed records primary key in a log file or in an error table. I have followed this approach many time for updating millions of records in the PROD database, as it is very safe approach.
I managed to integrate Liquibase into our Maven build to initialize a H2 inmemory database with a few enrys. Those rows have the primary key generated using a sequence table which works as expected (BigInt incremented values starting from 1).
My issue is that when i try to persist a new entity into that table from within a Junit integration test i get a "unique key constraint violation" because that new entity has the same primary key as the very first row inserted using the Liquibase changelog-xmls.
So the initialisation works perfectly fine as expected. The maven build uses the liquibase changelog-xmls
For now i just wipe the according tables completly before any integration tests with an own Runner... but that wont be a possibility in the furture. Its currently quite a chalange to investigate such issues since there is not yet much specific information on Liquibase available.
Update Workaround
While id prefer below answer using H2 brings up the problem that below changeset wont work because the required minValue is not supported.
<changeSet author="liquibase-docs" id="alterSequence-example">
<alterSequence
incrementBy="1"
maxValue="371717"
minValue="40"
ordered="true"
schemaName="public"
sequenceName="seq_id"/>
As a simple workaround i now just drop the existing sequence that was used to insert my testdata in a second changeSet:
<changeSet id="2" author="Me">
<dropSequence
sequenceName="SEQ_KEY_MY_TBL"/>
<createSequence
sequenceName="SEQ_KEY_MY_TBL"
incrementBy="1"
startValue="40"/>
</changeSet>
This way the values configured in the changelog-*.xml will be inserted using the sequence with an initial value of 1. I insert 30 rows so Keys 1-30 are used. After that the sequence gets dropped and recreated with a higher startValue. This way when persisting entities from within a Junit based integration Test the new entities will have primary keys starting from 40 and the previous unique constraint problem is solved.
Not H2 will probably soon release a version supporting minValue/maxValue since the according patch already exists.
Update:
Maybe we should mention this still is just a Workaround, anyone knows if H2 supports a Sequence with Liquibase that wont start over after DB-Init?
You should instruct liquibase to set the start value for those sequences to a value beyond those you have used for the entries you created. Liquibase has an alterSequence element for this. You can add such elements at the end of your current liquibase script.
I'm using Hibernate and a MySql server. I use multiple databases as "namespaces" (e.g. users, transactions, logging etc.).
So, I configued Hibernate to NOT connect to a particular database :
url = jdbc:mysql://127.0.0.1/
The databases where tables are located are defined in the hbm files through the catalog attribute :
<class name="com.myApp.entities.User" table="user" schema="" catalog="users"> ...
When I want to load some data, everything works fine and Hibernate seems to generate the expected SQL queries (by using the catalog prefix in the table names) e.g. :
select id from users.user
However, when I try to add a new record, Hibernate don't use the from [catalog].[table_name] syntax anymore. So I get a MySQL error 'No database selected'.
select max(id) from user
Hibernate is trying the get the future id to create a new record, but it doesn't specify in which database is located the table, it should be :
select max(id) from users.user
Why is Hibernate generating this invalid query ? Have someone ever experienced this problem ?
You need to specify the schema for the generator. See this question on SO for a more detailed answer.
I convert the DB from Oracle to MySQL.
I'm using both Java & Hibernate.
When I used oracle I had the following method that gave me a brand new and unused sequence value:
protected int getSequenceNextValue() {
Session session = sessionFactory.getCurrentSession();
Query query = session.createSQLQuery("select MY_SEQUENCE.NEXTVAL from DUAL");
return ((BigDecimal) query.uniqueResult()).intValueExact();
}
And I'm trying to refactor this method to work on MySQL DB.
I have a table in MySQL that I use as a sequence (through Hibernate):
create table MY_SEQUENCE(
next_val int(10) NOT NULL
);
Is there any thread safe way to get a new value from this table and in the same transction to increase it?
For most cases I use the Hibernate Generator to generate a new sequence using this table, but in several cases I need to do it manually.
The best solution for me will be a refactoring of the method above, in such way that threads that querying the table at the same time will not fail, but will wait for each other.
Thanks...
Have a look at the InnoDB table type and FOR UPDATE. An example similar to what you describe is in the MySQL manual here http://dev.mysql.com/doc/refman/5.0/en/innodb-locking-reads.html