com.atlassian.confluence.internal.upgrade.constraint.dedup

Class Deduper

java.lang.Object

com.atlassian.confluence.internal.upgrade.constraint.dedup.Deduper

public class Deduper
extends Object

Remove duplicated rows from a table before a unique constraint can be added. This is a 2-step process: 1. Query all duplicate rows 2. Use a suitable DedupeStrategy to delete those duplicates Note: 1. Tables with composite keys are not supported. 2. This implementation doesn't scale well because it puts everything in memory and process DMLs one at a time

Since:

7.1.0

Constructor Summary

Constructors

Constructor and Description
Deduper(org.hibernate.SessionFactory sessionFactory, com.atlassian.config.db.HibernateConfig hibernateConfig)

Method Summary

Modifier and Type	Method and Description
boolean	multipleNullsNotAllowed() Oracle and SQL Server behave differently from the rest: two rows that contain nulls for one or more key columns and the same combination of values for the other key columns violate a unique constraint.
long	removeDuplicates(String table, String idColumn, List<String> uniqueColumns, DedupeStrategy dedupeStrategy) Remove all duplicates of combination of uniqueColumns from table.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Deduper

public Deduper(org.hibernate.SessionFactory sessionFactory,
               com.atlassian.config.db.HibernateConfig hibernateConfig)

Method Detail

removeDuplicates

public long removeDuplicates(String table,
                             String idColumn,
                             List<String> uniqueColumns,
                             DedupeStrategy dedupeStrategy)
                      throws UpgradeException

Remove all duplicates of combination of uniqueColumns from table. Note: names of tables and columns are not to be quoted.

Returns:

number of duplicate rows removed

Throws:

UpgradeException

multipleNullsNotAllowed

public boolean multipleNullsNotAllowed()

Oracle and SQL Server behave differently from the rest: two rows that contain nulls for one or more key columns and the same combination of values for the other key columns violate a unique constraint.

Returns:

true if current SQL vendor doesn't accepts multiple null values in multi-column unique constraints