diff --git a/mllib/src/main/scala/org/apache/spark/mllib/linalg/distributed/RowMatrix.scala b/mllib/src/main/scala/org/apache/spark/mllib/linalg/distributed/RowMatrix.scala
index e76bc9fefff01..2e414a73be8e0 100644
--- a/mllib/src/main/scala/org/apache/spark/mllib/linalg/distributed/RowMatrix.scala
+++ b/mllib/src/main/scala/org/apache/spark/mllib/linalg/distributed/RowMatrix.scala
@@ -53,8 +53,14 @@ class RowMatrix(
   /** Gets or computes the number of columns. */
   override def numCols(): Long = {
     if (nCols <= 0) {
-      // Calling `first` will throw an exception if `rows` is empty.
-      nCols = rows.first().size
+      try {
+        // Calling `first` will throw an exception if `rows` is empty.
+        nCols = rows.first().size
+      } catch {
+        case err: UnsupportedOperationException =>
+          sys.error("Cannot determine the number of cols because it is not specified in the " +
+            "constructor and the rows RDD is empty.")
+      }
     }
     nCols
   }
@@ -293,6 +299,10 @@ class RowMatrix(
         (s1._1 + s2._1, s1._2 += s2._2)
     )
 
+    if (m <= 1) {
+      sys.error(s"RowMatrix.computeCovariance called on matrix with only $m rows." +
+        "  Cannot compute the covariance of a RowMatrix with <= 1 row.")
+    }
     updateNumRows(m)
 
     mean :/= m.toDouble
diff --git a/mllib/src/test/scala/org/apache/spark/mllib/stat/CorrelationSuite.scala b/mllib/src/test/scala/org/apache/spark/mllib/stat/CorrelationSuite.scala
index f7d5980c196fa..34548c86ebc14 100644
--- a/mllib/src/test/scala/org/apache/spark/mllib/stat/CorrelationSuite.scala
+++ b/mllib/src/test/scala/org/apache/spark/mllib/stat/CorrelationSuite.scala
@@ -42,8 +42,12 @@ class CorrelationSuite extends FunSuite with LocalSparkContext {
   test("corr(x, y) pearson, 1 value in data") {
     val x = sc.parallelize(Array(1.0))
     val y = sc.parallelize(Array(4.0))
-    assert(Statistics.corr(x, y, "pearson").isNaN)
-    assert(Statistics.corr(x, y, "spearman").isNaN)
+    intercept[RuntimeException] {
+      Statistics.corr(x, y, "pearson")
+    }
+    intercept[RuntimeException] {
+      Statistics.corr(x, y, "spearman")
+    }
   }
 
   test("corr(x, y) default, pearson") {