Manage the update order for queries in JPA - spring

Question

I am creating a simple kanban application as following, each kanban is made out of a sequence of stages and each stage have a level field to define its position. I want to be able to add, move and remove stages at will so I have to keep the level of each stage consistent, simple enough.
#Entity
#Table(name = "kanbans")
data class Kanban (
#Id
#GeneratedValue(strategy = GenerationType.IDENTITY)
#Column(name = "id", nullable = false)
var id: Int? = null,
#get:NotNull
#get:NotBlank
#Column(name = "name", nullable = false)
var name: String? = null,
#get:NotNull
#get:NotBlank
#Column(name = "description", nullable = false)
var description: String? = null,
#get:NotNull
#Column(name = "closed", nullable = false)
var closed: Boolean? = null,
#get:NotNull
#Column(name = "created_at", nullable = false)
var createdAt: LocalDateTime? = null,
#get:NotNull
#Column(name = "updated_at", nullable = false)
var updatedAt: LocalDateTime? = null,
)
#Entity
#Table(name = "stages")
data class Stage (
#Id
#GeneratedValue(strategy = GenerationType.IDENTITY)
#Column(name = "id", nullable = false)
var id: Int? = null,
#get:NotNull
#get:NotBlank
#Column(name = "name", nullable = false)
var name: String? = null,
#get:NotNull
#get:NotBlank
#Column(name = "description", nullable = false)
var description: String? = null,
#get:NotNull
#Column(name = "closed", nullable = false)
var closed: Boolean? = null,
#get:NotNull
#Column(name = "level", nullable = false)
var level: Int? = null,
#OneToMany(fetch = FetchType.LAZY, mappedBy = "stage")
var tasks: List<Task> = ArrayList(),
#get:NotNull
#ManyToOne(fetch = FetchType.LAZY)
#JoinColumn(name = "kanban_id", nullable = false)
var kanban: Kanban? = null,
#get:NotNull
#Column(name = "created_at", nullable = false)
var createdAt: LocalDateTime? = null,
#get:NotNull
#Column(name = "updated_at", nullable = false)
var updatedAt: LocalDateTime? = null,
)
When creating the first stage its always assigning its level at 0 and then when adding new ones the level will define the stage position at the list of stages. The problem is that when I try to update the previous existing stages to give place to the new one, the only way I found to make this work is to place a saveAndFlush call in a loop but I find it to be not a good ideia.
#Repository
interface StageRepository : JpaRepository<Stage, Int> {
fun findAllByKanbanAndLevelGreaterThanEqualOrderByLevelDesc(kanban: Kanban, level: Int): List<Stage>
#Modifying
#Transactional
#Query("UPDATE Stage s SET s.level = s.level + 1 WHERE s.kanban = :kanban AND s.level >= :level")
fun incrementLevelForKanbanStagesWhereLevelIsGreaterThan(kanban: Kanban, level: Int)
}
the incrementLevelForKanbanStagesWhereLevelIsGreaterThan method fails as the database have a unique constraint to level and kanban_id with the following error:
Caused by: org.postgresql.util.PSQLException: ERROR: duplicate key value violates unique constraint "stages_kanban_id_level_key"
Detalhe: Key (kanban_id, level)=(337, 1) already exists.
this is obviously happening because it is trying to update level 0 to level 1 before updating level 1 to level 2 and so I have tried:
#Transactional
#Query("UPDATE Stage s SET s.level = s.level + 1 WHERE s.kanban = :kanban AND s.level >= :level ORDER BY s.level DESC")
fun incrementLevelForKanbanStagesWhereLevelIsGreaterThan(kanban: Kanban, level: Int)
which does not compile,
#Service
#Transactional
class StageCrudService: CrudService<Stage, Int, StageRepository, StageValidationService>() {
#Throws(ValidationException::class)
override fun create(model: Stage): Stage {
prepareToCreate(model)
validationService.canSave(model)
incrementKanbanStageLevels(model)
return repository.save(model)
}
private fun prepareToCreate(model: Stage) {
val now = LocalDateTime.now()
val closed = model.closed ?: false
model.closed = closed
model.createdAt = now
model.updatedAt = now
model.level = model.level ?: 0
}
private fun incrementKanbanStageLevels(model: Stage) {
val level = model.level ?: 0
val stages = repository.findAllByKanbanAndLevelGreaterThanEqualOrderByLevelDesc(model.kanban!!, level)
stages.forEach { stage ->
stage.level = stage.level?.plus(1)
}
repository.saveAll(stages)
repository.flush()
}
}
and
private fun incrementKanbanStageLevels(model: Stage) {
val level = model.level ?: 0
val stages = repository.findAllByKanbanAndLevelGreaterThanEqualOrderByLevelDesc(model.kanban!!, level)
stages.forEach { stage ->
stage.level = stage.level?.plus(1)
repository.save(stage)
}
repository.flush()
}
but both fails the same way as the query. Now the question is:
Is there a better way to manage the update order for this kind of situation instead of doing:
private fun incrementKanbanStageLevels(model: Stage) {
val level = model.level ?: 0
val stages = repository.findAllByKanbanAndLevelGreaterThanEqualOrderByLevelDesc(model.kanban!!, level)
stages.forEach { stage ->
stage.level = stage.level?.plus(1)
repository.saveAndFlush(stage)
}
}

Answer 1

It seems to me that you are possibly trying to implement something that can be managed for you via the JPA #OrderColumn annotation:
https://docs.oracle.com/javaee/7/api/javax/persistence/OrderColumn.html
Specifies a column that is used to maintain the persistent order of a
list. The persistence provider is responsible for maintaining the
order upon retrieval and in the database. The persistence provider is
responsible for updating the ordering upon flushing to the database to
reflect any insertion, deletion, or reordering affecting the list.
To use this you would need to make the relationship bi-directional and the level should be maintained by your JPA provider as items are added to and removed from the list
#Entity
#Table(name = "kanbans")
data class Kanban (
.....
#get:NotNull
#get:NotBlank
#OrderColumn(name = "level")
#OneToMany(fetch = FetchType.LAZY, mappedBy = "kanban")
var stage: List<Stage> = ArrayList()
.....
}
So you can then remove and add items (at any position) and the sequence will be maintained for you.