only look at the first values for first()

Since we set up the aggregate queries so that iterator is ordered,
we only need to look at the first value. This cuts about 10 seconds
off a large single series query running first.

influxql/functions.go

@@ -935,28 +935,20 @@ type firstLastMapOutput struct {
 }
 
 // MapFirst collects the values to pass to the reducer
+// This function assumes time ordered input
 func MapFirst(itr Iterator) interface{} {
-	out := &firstLastMapOutput{}
-	pointsYielded := false
-
-	for k, v := itr.Next(); k != -1; k, v = itr.Next() {
-		// Initialize first
-		if !pointsYielded {
-			out.Time = k
-			out.Val = v
-			pointsYielded = true
-		}
-		if k < out.Time {
-			out.Time = k
-			out.Val = v
-		} else if k == out.Time && greaterThan(v, out.Val) {
-			out.Val = v
-		}
+	k, v := itr.Next()
+	if k == -1 {
+		return nil
 	}
-	if pointsYielded {
-		return out
-	}
-	return nil
+	nextk, nextv := itr.Next()
+	for nextk == k {
+		if greaterThan(nextv, v) {
+			v = nextv
+		}
+		nextk, nextv = itr.Next()
+	} 
+	return &firstLastMapOutput{k, v}
 }
 
 // ReduceFirst computes the first of value.