Netflix Astyanax in scala - scala

Question

I got this code from https://github.com/Netflix/astyanax/wiki/Cql-and-cql3
I'm new to scala and my problem is to convert this for loop to a scala one.
Could you please help ?
result = keyspace
.prepareQuery(CQL3_CF)
.withCql("SELECT * FROM employees WHERE empId='111';")
.execute();
for (Row<Integer, String> row : result.getResult().getRows()) {
LOG.info("CQL Key: " + row.getKey());
ColumnList<String> columns = row.getColumns();
LOG.info(" empid : " + columns.getIntegerValue("empid", null));
LOG.info(" deptid : " + columns.getIntegerValue("deptid", null));
LOG.info(" first_name : " + columns.getStringValue ("first_name", null));
LOG.info(" last_name : " + columns.getStringValue ("last_name", null));
}
Many thanks

Answer 1

One way to convert this code into a scala based looping construct would be as follows:
import scala.collection.JavaConversions._
val result = keyspace
.prepareQuery(CQL3_CF)
.withCql("SELECT * FROM employees WHERE empId='111';")
.execute();
result.getResult().getRows() foreach { row =>
LOG.info("CQL Key: " + row.getKey())
val columns = row.getColumns()
LOG.info(" empid : " + columns.getIntegerValue("empid", null))
LOG.info(" deptid : " + columns.getIntegerValue("deptid", null))
LOG.info(" first_name : " + columns.getStringValue ("first_name", null))
LOG.info(" last_name : " + columns.getStringValue ("last_name", null))
}
By importing in JavaConversions._, we gain access to an implicit conversion that will convert a Java Iterable (which the Rows object is) to a scala.collection.Iterable, thus allowing you to use the foreach looping construct.
Now, this code is syntactically sound, but it's not really good Scala code. It's not very functional in that the loop itself does not return anything. It also has some messy logic dealing with nulls that should probably use Options instead. A more functional example of the solution, using map to map the result data into a simple case class could look like this:
import scala.collection.JavaConversions._
val result = keyspace
.prepareQuery(CQL3_CF)
.withCql("SELECT * FROM employees WHERE empId='111';")
.execute();
case class Employee(id:Option[Int], depId:Option[Int],
firstName:Option[String], lastName:Option[String])
def optFor[T](cl:ColumnList[String], func:(ColumnList[String] => T)):Option[T] = {
func(cl) match{
case null => None
case nonnull => Some(nonnull)
}
}
val employees = result.getResult().getRows() map { row =>
LOG.info("CQL Key: " + row.getKey())
val cl = row.getColumns()
val employee = Employee(optFor(cl, _.getIntegerValue("empid", null)),
optFor(cl, _.getIntegerValue("deptid", null)),
optFor(cl, _.getStringValue("first_name", null)),
optFor(cl, _.getStringValue("last_name", null)))
LOG.info(employee)
employee
}
There's probably a more elegant way to handle the null to Option conversion (via implicits perhaps), but this works as well. When done with the map operation, you will have an scala.collection.Iterable of Employee instances that you can then then perhaps return out to the UI for display.