Hibernate diffChangeLog producing change sets that exist already in the table - spring

Question

I am using spring boot with hibernate 5.3.8. After I run the first schema I generate a diffChangeLog and its dropping and creating the same index. When I look at the collection_days table a common_collection_search exists
How can I mark it as completed so it no longer appears in the diffChangeLog. I only want generate a changelog with the newest changes so I can run it and update the database
collection_days table
index_name|index_algorithm|is_unique|column_name|condition
common_collection_search|BTREE|f|waste_generator_id,refuse_type|NULL
This is from my master schema that I ran on a fresh database
- changeSet:
id: 1572649828026-46
author: comp (generated)
changes:
- createIndex:
columns:
- column:
name: wg_id
- column:
name: type
indexName: common_collection_search
tableName: collection_days
diffChangeLog results
- changeSet:
id: 1615578168770-16
author: comp (generated)
changes:
- dropIndex:
indexName: common_collection_search
tableName: collection_days
- changeSet:
id: 1615578168770-17
author: comp (generated)
changes:
- createIndex:
columns:
- column:
name: wg_id
- column:
name: type
indexName: common_collection_search
tableName: collection_days
CollectionDay entity
#Data
#Entity
#Table(name = "collection_days", uniqueConstraints = {
#UniqueConstraint(columnNames = { "wg_id", "type", "day_of_week" }) }, indexes = {
#Index(name = "common_collection_search", columnList = "wg_id,type") })
public class CollectionDay implements Serializable {
#Column(name = "wg_id",updatable = false,insertable = false)
private Long wgId;
#Column(name = "type")
#Enumerated(EnumType.STRING)
private ESEnums.type type;
build.gradle
compile "org.hibernate:hibernate-core"
compile "org.hibernate:hibernate-entitymanager"
compile "org.hibernate:hibernate-envers"
compile "org.springframework.boot:spring-boot-starter-data-jpa"
compile "com.zaxxer:HikariCP"
compile "org.hibernate:hibernate-ehcache"
compile 'com.fasterxml.jackson.datatype:jackson-datatype-hibernate5'
compile "org.postgresql:postgresql"
compile "org.liquibase:liquibase-core"
compile "org.hibernate:hibernate-validator:6.1.6.Final"
liquibaseRuntime 'org.liquibase.ext:liquibase-hibernate5:3.8'
liquibaseRuntime sourceSets.main.output
liquibaseRuntime sourceSets.main.compileClasspath
liquibase {
activities {
main {
driver 'org.postgresql.Driver'
url 'jdbc:postgresql://localhost:54323/db2?stringtype=unspecified'
username ''
password ''
changeLogFile "src/main/resources/liquibase/migrations/changelog.yaml"
classpath 'src/main/java'
referenceUrl 'hibernate:spring:com.project.domain.entities?dialect=org.hibernate.dialect.PostgreSQL82Dialect&hibernate.physical_naming_strategy=org.springframework.boot.orm.jpa.hibernate.SpringPhysicalNamingStrategy&hibernate.implicit_naming_strategy=org.springframework.boot.orm.jpa.hibernate.SpringImplicitNamingStrategy'
}
}
}

Answer 1

I think the problem is that PostgreSQL is smart enough to realize that the index is unnecessary and doesn't create it in the way you expect it. Your unique constraint already covers the columns of the index and since it is backed by an index, it will just tell you that it won't create your index, because it is already covered.