KotliQuery is a handy RDB client library for Kotlin developers! The design is highly inspired by ScalikeJDBC, which is a proven database library in Scala. The priorities in this project are:

• Less learning time
• No breaking changes in releases
• No additional complexity on top of JDBC

This library simply mitigates some pain points of the JDBC but our goal is not to completely encapsulate it.

The quickest way to try this library out would be to start with a simple Gradle project. You can find some examples here.

apply plugin: 'kotlin'

buildscript {
    ext.kotlin_version = '1.7.20'
    repositories {
        mavenCentral()
    }
    dependencies {
        classpath "org.jetbrains.kotlin:kotlin-gradle-plugin:$kotlin_version"
    }
}
repositories {
    mavenCentral()
}
dependencies {
    implementation "org.jetbrains.kotlin:kotlin-stdlib:$kotlin_version"
    implementation 'com.github.seratch:kotliquery:1.9.0'
    implementation 'com.h2database:h2:2.1.214'
}

KotliQuery is much more easy-to-use than you expect. After just reading this short section, you will have learnt enough.

First thing you do is to create a Session object, which is a thin wrapper of java.sql.Connection instance. With this object, you can run queries using an established database connection.

import kotliquery.*

val session = sessionOf("jdbc:h2:mem:hello", "user", "pass") 

For production-grade applications, utilizing a connection pool library for better performance and resource management is highly recommended. KotliQuery provides an out-of-the-box solution that leverages HikariCP, which is a widely accepted connection pool library.

HikariCP.default("jdbc:h2:mem:hello", "user", "pass")

sessionOf(HikariCP.dataSource()).use { session ->
   // working with the session
}

You can use a session for executing both DDLs and DMLs. The asExecute method if a query object sets the underlying JDBC Statement method to execute.

session.run(queryOf("""
  create table members (
    id serial not null primary key,
    name varchar(64),
    created_at timestamp not null
  )
""").asExecute) // returns Boolean

Using asUpdate is an appropriate way to perform insert/update/delete statements. This method sets the underlying JDBC Statement method to executeUpdate.

val insertQuery: String = "insert into members (name,  created_at) values (?, ?)"

session.run(queryOf(insertQuery, "Alice", Date()).asUpdate) // returns effected row count
session.run(queryOf(insertQuery, "Bob", Date()).asUpdate)

Now that you've got a database table named members, it's time to run your first SQL statement with this library! To build a callable SQL executor, your code follows the three steps for it:

• Use queryOf factory method with a query statement and its parameters to create a new Query object
• Use #map method to attache a result extracting function ((Row) -> A) to the Query object
• Specify the response type (asList/asSingle) for the result

The following query returns a list of all member's IDs. In this line, the SQL statement is not yet executed. Also, this object allIdsQuery does not have any state. This means that you can reuse th object multiple times.

val allIdsQuery = queryOf("select id from members").map { row -> row.int("id") }.asList

With a valid session object, you can perform the SQL statement. The type of returned ids would be safely determined by Kotlin compiler.

val allIds: List<Int> = session.run(allIdsQuery)

As you see, the extractor function is greatly flexible. You can define functions with any return type. All you need to do is to implement a function that extracts values from JDBC ResultSet interator and map them into a single expected type value. Here is a complete example:

data class Member(
  val id: Int,
  val name: String?,
  val createdAt: java.time.ZonedDateTime)

val toMember: (Row) -> Member = { row -> 
  Member(
    row.int("id"), 
    row.stringOrNull("name"), 
    row.zonedDateTime("created_at")
  )
}

val allMembersQuery = queryOf("select id, name, created_at from members").map(toMember).asList
val allMembers: List<Member> = session.run(allMembersQuery)

If you are sure that a query can return zero or one row, asSingle returns an optional single value as below:

val aliceQuery = queryOf("select id, name, created_at from members where name = ?", "Alice").map(toMember).asSingle
val alice: Member? = session.run(aliceQuery)

Technically, it's also possible to use asSingle along with an SQL statement returning multiple rows. With the default setting, the result data extraction returns only the first row in the results and skips the rest. In other words, KotliQuery silently ignores the inefficiency and the potential misbehavior. If you prefer detection by an error in this scenario, you can pass strict flag to Session initializer. With strict set to true, the query execution throws an exception if it detects multiple rows for asSingle.

// Session object constructor
val session = Session(HikariCP.dataSource(), strict = true)

// an auto-closing code block for session
sessionOf(HikariCP.dataSource(), strict = true).use { session ->

}

An alternative way to bind parameters is to use named parameters that start with : in the statement string. Note that, with this feature, KotliQuery still uses a prepared statement internally and your query execution is safe from SQL injection. The parameter parts like :name and :age in the following example query won't be just replaced as string values.

queryOf(
  """
  select id, name, created_at 
  from members 
  where (name = :name) and (age = :age)
  """, 
  mapOf("name" to "Alice", "age" to 20)
)

Performance-wise, the named parameter syntax can be slightly slower for parsing the statement plus a tiny bit more memory-consuming. But for most use case, the overhead should be ignorable. If you would like to make your SQL statements more readable and/or if your query has to repeat the same parameter in a query, using named query parameters should improve your productivity and the maintainability of the query a lot.

You can specify the Java type for each parameter in the following way. Passing the class Parameter helps KotliQuery properly determine the type to bind for each parameter in queries.

val param = Parameter(param, String::class.java)
queryOf(
  """
  select id, name
  from members 
  where ? is null or ? = name
  """,
  param,
  param
)

As a handier way, you can use the following helper method.

queryOf(
  """
  select id, name 
  from members 
  where ? is null or ? = name
  """, 
  null.param<String>(),
  null.param<String>()
)

This functionality is particularly useful in the situations like the ones dsecribed here.

The #forEach allows you to work with each row with less memory consumption. With this way, your application code does not need to load all the query result data in memory at once. This feature is greatly useful when you load a large number of rows from a database table by a single query.

session.forEach(queryOf("select id from members")) { row ->
  // working with large data set
})

Running queries in a transaction is of course supported! The Session object provides a way to start a transaction in a certain code block.

session.transaction { tx ->
  // begin
  tx.run(queryOf("insert into members (name,  created_at) values (?, ?)", "Alice", Date()).asUpdate)
}
// commit

session.transaction { tx ->
  // begin
  tx.run(queryOf("update members set name = ? where id = ?", "Chris", 1).asUpdate)
  throw RuntimeException() // rollback
}

As this library is a bit opinionated, transactions are available only with a code block. We intentionally do not support begin / commit methods. If you would like to manually manage the state of a transaction for some reason, you can use session.connection.commit() / session.connection.rollback() for it.

• Pro Kotlin Web Apps from Scratch by August Lilleaas: uses KotliQuery to demonstrate how to efficiently run SQL queries

The MIT License Copyright (c) 2015 - Kazuhiro Sera