Why is this throwing a

JdbcSQLException: Database is already closed (to disable automatic closing at VM shutdown, add ";DB_CLOSE_ON_EXIT=FALSE" to the db URL)

while close on exit = false is already set?

@Before
    fun setupDatabase() {
        Database.connect("jdbc:h2:mem:test;DB_CLOSE_ON_EXIT=FALSE;DB_CLOSE_DELAY=-1", "org.h2.Driver")

        // create tables
    }

    @After
    fun tearDownDatabase() {
        transaction {
            exec("SHUTDOWN;")
        }
    }

Hi. I was looking for a library like Exposed but found as well @orangy’s squash at https://github.com/orangy/squash. What's the relation between these projects?

For some reason when calling a field with text type in H2,

org.h2.jdbc.JdbcSQLException: The object is already closed

is thrown. At:

val map = transaction { loggedRequest.parameters } // throws exception

val map = transaction { loggedRequest.method } // doesn't throw exception

For:

object ApiLogs : IntIdTable() {
    ...
    val parameters = text("parameters")
    val method = varchar("method", 191)
    ...
}

class ApiLog(id: EntityID<Int>) : IntEntity(id) {
    companion object : IntEntityClass<ApiLog>(ApiLogs)

    ...
    var parameters by ApiLogs.parameters
    var method by ApiLogs.method
    ...
}

If I have an existing postgres schema managed by a python framework, but would like my api to do CRUD on only part of the schema, can I just declare part of it in exposed at use only the part I need? I just need CRUD, not that it should change the structure at all...

For the DAO method, when do updates occur? The wiki says '"Flushing" values to the database occurs at the end of the transaction or before next "select *" from the database.' What happens if I change my object outside of a transaction statement?

Also, coroutine support could still be for non async using a thread pool, at least until there would be an async solution... otherwise anyone using this with Ktor will have to surround everything with

async { }.await()

and make their own thread pools... I have a coroutine utils package in most of my projects for such things that I keep having to copy... it would be great if it could be integrated into the lib...

Hi, why am I getting this error:

[14:31:02 ERROR]: [Exposed] mysql 5.5 doesn't support expression '(CURRENT_TIMESTAMP)' as default value. Column will be created with NULL marker.

. I'm using MariaDB 10.1.26.

i'm working with DAOs. i have a table that has a foreign key /

optionalReferencedOn

, is it possible in a query to access the referenced object? (for example

User

has a reference to

City

, and

City

has a column

Inhabitants

and i want to get all users living in a

city

with at least 5k inhabitants) (typing it, it seems like it would be a better approach from "the other side" (looking for cities with inhabitants and getting all

users

that have a reference to those cities, but not sure how i'd do that either atm)

in case I would like to use a column with whole number that is bigger than

long

, what are my options?

decimal

?

Is there a way to retreive only an sql statement w/o running it (to be executed w/ another library...)?

Hi. I feel like i asked this already long time ago, but: how do i create multi-column PK (i.e. for many-to-many join tables)? do i just call primaryKey() on all required columns? ... It does.

Hi! I'm getting this very annoying exception:

Caused by: java.sql.SQLException: Multiple primary key defined
Query is: ALTER TABLE Clans MODIFY COLUMN id INT AUTO_INCREMENT PRIMARY KEY

Does anyone know why this is happening? This happens everytime I run

SchemaUtils.createMissingTablesAndColumns(Clans)

with Clans or with any other table.

@tapac - after upgrading to 0.10.1 the following do not compile:

transaction(statement)

In your case you should use

insertAndGetId

When you write

insert[id]

you try to get value related to "expression" from resultSet. It can be null if, for example, it wasn't inserted, that's why column type (not-nullable) doesn't make sense.

This is what I went for, is there a cleaner or more idiomatic way to do it?

return transaction {
           //Check if a user with this device id exists       
            val userIdEntity = Users.select {
                Users.deviceId eq deviceId
            }.map { it[Users.id] }.firstOrNull()

            //Assign the user id 
            val userId =
                    if(userIdEntity != null) {
                        userIdEntity.value
                    } else
                    {
                        Users.insertAndGetId {
                        //...
                        }.value
                    }
        }

Is there a way to select isNull() ?

I am able to achieve a connection pool to my database using SpringTransactionManager(dataSource) . But as my dao layer use transaction { } blocks, I get the IllegalStateException: Already value for key bound to thread [main]

Hello there. Real noob here with kotlin and exposed. I'm trying to run a simple join but clearly I'm missing something obvious so I was trying to recreate what I read on the wiki. AccessToken.idStaff and Staff.id are both int. The exception says that "Cannot join as there is no matching primary key/ foreign key pair and constraint missing". Isn't that "select" the condition for the join?

(AccessToken innerJoin Staff).slice(Staff.id, Staff.userName).select {AccessToken.idStaff eq Staff.id}

Is there any way to map them to Inspection Entity object directly?

I get this error:

... don't have client default values. DB defaults don't support in batch inserts

why is that not supported?

I am not sure why does it have an extra phone in the create statement.

Hi! I'm starting with kotlin+exposed, how does the Database object got passed to transaction {} method?

Database.connect("<url>", driver = "org.postgresql.Driver", user = "..", password = "...")
transaction {

and which is the common approach to initialise Database, so I don't create it in each service method? Thanks in advance

Is there something like Django's "Q" object?

this doesn't compile anymore after updating:

ResultRow.get(Max(Table.id, IntegerColumnType()))

Hello, I'm new to Kotlin and Exposed, and I'm having some difficulty with the DAO api. I'm trying to create a function that will have a where clause that is different depending on runtime user input. I don't appear to be able to use any if blocks inside the transaction block. I'm trying to do something like this:

fun findX(xGt: BigDecimal?, xLt: BigDecimal?): List<DaoClass> {
    return transaction {
        DaoClass.find {
            var condition = TableClass.any()
            if (xGt != null) {
                condition = condition.and(TableClass.x.greater(xGt))
            }
            if (xLt != null) {
                condition = condition.and(TableClass.x.less(xGt))
            }

            condition
        }.toList()
    }
}

Is this possible?

can DAO do everything DSL can?

Is there a way to specify the encoding for certain columns? Specifically, I’m looking to tell MySQL that some columns are utf8mb4. In case it matters, we’re using the DAO approach.

Rather than using the implicit Orgs id, I want to reference on the Orgs (and Users) uid, which is a unique string. What’s the proper syntax for defining OrgMembership in that case? IntelliJ complains that

val org by Org referencedOn OrgMemberships.orgId

has a type mismatch, where it expects a

Column<EntityId<Int>>

but gets a

Column<EntityId<String>>

. How do I tell OrgMembership to use that string?