Grouping into an Array with Quill Spark

Example.scala

case class Address(street: String, zip: Int, personName: String)
case class Person(name: String, addresses: List[String])

/*
In some Data-Models used with Spark, you may want to embed arrays into an object.
You may then want to be able to set array-properties of this kind of object when
a group-by is done. Normally this would be done in Spark using the following way
*/
val addresses: Dataset[Address] = ???
val people = addresses.groupByKey(a => a.personName).mapGroups { case (name, addresses) =>
  Person(name, addresses.map(_.street).toList)
}

/*
If you are using Quill Queries however, this has to be done a little bit differently because all Quill-Spark 
queries need to be converted to SQL before they are run. The question then becomes, what kind of SQL Aggregator
to use to achieve the equivalent functionality. Fortunately, Spark has an aggregator called `collect_list` 
which does this.
*/
val addressesQuery: Dataset[Address] = liftQuery(addresses)
val peopleQuery: Query[Person] = quote {
  addressesQuery
    .groupBy(a => a.personName)
    .map { case (name, addresses) =>
      Person(name, infix"collect_list(${addresses.map(_.street)})".pure.as[List[String]])
    }
}
/*
This creates a Query that looks like the following:
SELECT 
  a.name,
  collect_list(a.street)
FROM Addresses a GROUP BY a.name
*/


/*
Alternatively, the list that is collected can be a list of objects.
*/
case class Address(street: String, zip: Int, personName: String)
case class Person(name: String, addresses: List[Address])

val addressesQuery: Dataset[Address] = liftQuery(addresses)
val peopleQuery: Query[Person] = quote {
  addressesQuery
    .groupBy(a => a.personName)
    .map { case (name, addresses) =>
      Person(name, infix"collect_list(${addresses})".pure.as[List[Address]])
    }
}

/*
This creates a Query that looks like the following:
SELECT 
  a.name,
  collect_list(a.*)
FROM Addresses a GROUP BY a.name
*/