1. 技术文章
  2. Unsuccessful: alter table XXX drop constraint YYY in Hibernate/JPA/HSQLDB standalone

Unsuccessful: alter table XXX drop constraint YYY in Hibernate/JPA/HSQLDB standalone

由 主宰稳场 提交于 2019-11-27 17:31:05

You can ignore these error. Combination of create-drop and in-memory database produces these for every database object it tries to drop. Reason being that there is not any database objects to remove - DROP statements are executed against empty database.

Also with normal permanent database such a errors do come, because Hibernate does not figure out before executing DROP statements does added object exist in database or is it new.

This solution worked for me, as opposed to the other solution that's given. Apparently mileages vary.

This was my exact error:

HHH000389: Unsuccessful: alter table ... drop constraint FK_g1uebn6mqk9qiaw45vnacmyo2 if exists
Table "..." not found; SQL statement: ...

This is my solution, overriding the H2 dialect:

package com.totaalsoftware.incidentmanager;

import org.hibernate.dialect.H2Dialect;

/**
 * Workaround.
 * 
 * @see https://hibernate.atlassian.net/browse/hhh-7002
 * 
 */
public class ImprovedH2Dialect extends H2Dialect {
    @Override
    public String getDropSequenceString(String sequenceName) {
        // Adding the "if exists" clause to avoid warnings
        return "drop sequence if exists " + sequenceName;
    }

    @Override
    public boolean dropConstraints() {
        // We don't need to drop constraints before dropping tables, that just
        // leads to error messages about missing tables when we don't have a
        // schema in the database
        return false;
    }
}

The Solution @Sander provided above works for MYSQL too. Just extend MySQL5InnoDBDialect instead like below:

import org.hibernate.dialect.MySQL5InnoDBDialect;

public class ImprovedMySQLDialect extends MySQL5InnoDBDialect {
    @Override
    public String getDropSequenceString(String sequenceName) {
        // Adding the "if exists" clause to avoid warnings
        return "drop sequence if exists " + sequenceName;
    }

    @Override
    public boolean dropConstraints() {
        // We don't need to drop constraints before dropping tables, that just leads to error
        // messages about missing tables when we don't have a schema in the database
        return false;
    }
}

Then in your datasource file change the following line:

dialect = org.hibernate.dialect.MySQL5InnoDBDialect

to

dialect = my.package.name.ImprovedMySQLDialect
10GritSandpaper

Just set dbCreate="update", and the errors go away immediately.

See: https://stackoverflow.com/a/31257468/715608

The annoying error messages became more obnoxious stack traces at the start of every test using an in-memory database with HSQLDB and hbm2ddl.auto=create-drop.

The response to an HSQLDB bug report suggested the use of DROP TABLE ... CASCADE rather than DROP CONSTRAINT IF EXISTS. Unfortunately the HSQLDB syntax for drop table is DROP TABLE <table> [IF EXISTS] [RESTRICT | CASCADE]; Hibernate Dialect does not provide an easy mechanism for a Dialect to specify CASCADE following the final IF EXISTS clause. I wrote a bug for this limitation.

However, I was able to overcome the issue by creating a custom Dialect as follows:

public class HsqlDialectReplacement extends HSQLDialect {

  @Override
  public String getDropTableString( String tableName ) {
    // Append CASCADE to formatted DROP TABLE string
    final String superDrop = super.getDropTableString( tableName );
    return superDrop + " cascade";
  }

  @Override
  public boolean dropConstraints() {
      // Do not explicitly drop constraints, use DROP TABLE ... CASCADE
      return false;
  }

  @Override
  public Exporter<Table> getTableExporter() {
    // Must override TableExporter because it also formulates DROP TABLE strings
    synchronized( this ) {
      if( null == overrideExporter ) {
        overrideExporter = new HsqlExporter( super.getTableExporter() );
      }
    }

    return overrideExporter;
  }

  private Exporter<Table> overrideExporter = null;

  private static class HsqlExporter implements Exporter<Table> {
    HsqlExporter( Exporter<Table> impl ) {
      this.impl = impl;
    }

    @Override
    public String[] getSqlCreateStrings( Table exportable, Metadata metadata ) {
      return impl.getSqlCreateStrings( exportable, metadata );
    }

    @Override
    public String[] getSqlDropStrings( Table exportable, Metadata metadata ) {
      final String[] implDrop = impl.getSqlDropStrings( exportable, metadata );
      final String[] dropStrings = new String[implDrop.length];
      for( int i=0; i<implDrop.length; ++i ) {
        dropStrings[i] = implDrop[i] + " cascade";
      }
      return dropStrings;
    }

    private final Exporter<Table> impl;
  };
}

We faced same issue while trying to create a simple login window in Spring. There were two tables user and role both with primary key id type of BIGINT. These we mapped (Many-to-Many) into another table user_roles with two columns user_id and role_id as foreign keys.

The problem was the role_id column in user_roles table, it was of int type and not compatible for the foreign key to role. When the type was modified to BIGINT errors were rectified.

标签
java
hibernate
jpa
hsqldb
in-memory
易学教程内所有资源均来自网络或用户发布的内容，如有违反法律规定的内容欢迎反馈！
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!