SpringBoot Spring Naming Strategy works unexpectedly for column with single char in the end

Hello Spring Data team! Off topic: You are awesome!

We are using spring-data-jpa with spring naming strategy and we found bug when its generating a column name for entity with property ends with single char, like for example: - myTypeA which I'm expecting should result into my_type_a - or myTypeB => my_type_b

spring:
  jpa:
    hibernate:
      naming:
        implicit-strategy: org.springframework.boot.orm.jpa.hibernate.SpringImplicitNamingStrategy
        physical-strategy: org.springframework.boot.orm.jpa.hibernate.SpringPhysicalNamingStrategy

Spring Data JPA Entity

@Entity
data class MyEntity(

    @Id
    @GeneratedValue(strategy = IDENTITY)
    val id: Long? = null,

    // @Column(name = "my_type_a") // when @Column annotation is commented resulted in my_typea
    val myTypeA: String = "",

    // @Column(name = "my_type_b") // when @Column annotation is commented resulted in my_typeb
    val myTypeB: String = "",
)

I'm expecting that myTypeA and myTypeB will be resulted into my_type_a and my_type_b, but actual values are my_typea and my_typeb

If you look at org.springframework.boot.orm.jpa.hibernate.SpringPhysicalNamingStrategy#apply method, you can see: SpringBoot Spring Naming Strategy works unexpectedly for column with single char in the end

that last char is simply ignoring

isUnderscoreRequired(builder.charAt(i - 1), builder.charAt(i), builder.charAt(i + 1))

Environment: - Spring Boot 2.5.9 - JDK 11

Regards, Maksim

Comment From: snicoll

@daggerok thanks for the report. Unfortunately, the naming strategies have moved to the Hibernate project (#27352) so we don't have the freedom to make a change that would apply consistently. Could you please report that against the Hibernate project? If you do so, adding a link to the created issue here would help us track the outcome. Thanks!