Question

If I have the following code:

var A = Array[Array[Double]]()    // where A becomes an MxP matrix
var B = Array[Array[Double]]()    // where B becomes an NxP matrix

What are some efficient ways to append one matrix to the other, resulting in a single matrix, as the following pseudocode would suggest?

val C = A append B    // where C is a (M+N)xP matrix

Obviously, one of the dimensions (in this case P) is held constant.

EDIT: So far, both of the provided solutions are growing in the second dimension. I am trying to hold the second dimension fixed.

Answer 1

Functional, but not as performant as the imperative alternative would be:

scala> val a = Array.tabulate(2, 3)((_, _) => (math.random * 100).toInt)
a: Array[Array[Int]] = Array(Array(52, 61, 58), Array(35, 69, 39))

scala> val b = Array.tabulate(2, 4)((_, _) => (math.random * 100).toInt)
b: Array[Array[Int]] = Array(Array(51, 54, 87, 10), Array(52, 76, 18, 85))

scala> (a, b).zipped.map(_ ++ _)
res0: Array[Array[Int]] = Array(Array(52, 61, 58, 51, 54, 87, 10), Array(35, 69, 39, 52, 76, 18, 85))

---

(In reply to the comment…)

Holding the second dimension fixed:

scala> val x = Array.tabulate(3, 2)((_, _) => (math.random * 100).toInt)
x: Array[Array[Int]] = Array(Array(13, 26), Array(96, 6), Array(68, 58))

scala> val y = Array.tabulate(2, 2)((_, _) => (math.random * 100).toInt)
y: Array[Array[Int]] = Array(Array(82, 5), Array(0, 76))

scala> x ++ y
res1: Array[Array[Int]] = Array(Array(13, 26), Array(96, 6), Array(68, 58), Array(82, 5), Array(0, 76))