app/vmselect/promql: properly handle (a op b) default N if (a op b) returns NaN series

The result should be a series with `N` values and `a op b` labels. Previously such series has been removed from the result.

app/vmselect/promql/binary_op.go

@@ -100,9 +100,8 @@ func newBinaryOpFunc(bf func(left, right float64, isBool bool) float64) binaryOp
 				dstValues[j] = bf(a, b, isBool)
 			}
 		}
-		// Optimization: remove time series containing only NaNs.
-		// This is quite common after applying filters like `q > 0`.
-		dst = removeNaNs(dst)
+		// Do not remove time series containing only NaNs, since then the `(foo op bar) default N`
+		// won't work as expected if `(foo op bar)` results to NaN series.
 		return dst, nil
 	}
 }

app/vmselect/promql/exec_test.go

@@ -1054,6 +1054,21 @@ func TestExecSuccess(t *testing.T) {
 		resultExpected := []netstorage.Result{r}
 		f(q, resultExpected)
 	})
+	t.Run("default_for_nan_series", func(t *testing.T) {
+		t.Parallel()
+		q := `label_set(0, "foo", "bar")/0 default 7`
+		r := netstorage.Result{
+			MetricName: metricNameExpected,
+			Values:     []float64{7, 7, 7, 7, 7, 7},
+			Timestamps: timestampsExpected,
+		}
+		r.MetricName.Tags = []storage.Tag{{
+			Key:   []byte("foo"),
+			Value: []byte("bar"),
+		}}
+		resultExpected := []netstorage.Result{r}
+		f(q, resultExpected)
+	})
 	t.Run(`alias()`, func(t *testing.T) {
 		t.Parallel()
 		q := `alias(time(), "foobar")`
@@ -1701,10 +1716,12 @@ func TestExecSuccess(t *testing.T) {
 			Values:     []float64{1123.456, 1323.456, 1523.456, 1723.456, 1923.456, 2123.456},
 			Timestamps: timestampsExpected,
 		}
-		r2.MetricName.Tags = []storage.Tag{{
-			Key:   []byte("foo"),
-			Value: []byte("123.456"),
-		}}
+		r2.MetricName.Tags = []storage.Tag{
+			{
+				Key:   []byte("foo"),
+				Value: []byte("123.456"),
+			},
+		}
 		resultExpected := []netstorage.Result{r1, r2}
 		f(q, resultExpected)
 	})
@@ -6961,19 +6978,21 @@ func testMetricNamesEqual(t *testing.T, mn, mnExpected *storage.MetricName, pos
 		t.Fatalf(`unexpected projectID; got %d; want %d`, mn.ProjectID, mnExpected.ProjectID)
 	}
 	if string(mn.MetricGroup) != string(mnExpected.MetricGroup) {
-		t.Fatalf(`unexpected MetricGroup at #%d; got %q; want %q`, pos, mn.MetricGroup, mnExpected.MetricGroup)
+		t.Fatalf(`unexpected MetricGroup at #%d; got %q; want %q; metricGot=%s, metricExpected=%s`,
+			pos, mn.MetricGroup, mnExpected.MetricGroup, mn.String(), mnExpected.String())
 	}
 	if len(mn.Tags) != len(mnExpected.Tags) {
-		t.Fatalf(`unexpected tags count at #%d; got %d; want %d`, pos, len(mn.Tags), len(mnExpected.Tags))
+		t.Fatalf(`unexpected tags count at #%d; got %d; want %d; metricGot=%s, metricExpected=%s`, pos, len(mn.Tags), len(mnExpected.Tags), mn.String(), mnExpected.String())
 	}
 	for i := range mn.Tags {
 		tag := &mn.Tags[i]
 		tagExpected := &mnExpected.Tags[i]
 		if string(tag.Key) != string(tagExpected.Key) {
-			t.Fatalf(`unexpected tag key at #%d,%d; got %q; want %q`, pos, i, tag.Key, tagExpected.Key)
+			t.Fatalf(`unexpected tag key at #%d,%d; got %q; want %q; metricGot=%s, metricExpected=%s`, pos, i, tag.Key, tagExpected.Key, mn.String(), mnExpected.String())
 		}
 		if string(tag.Value) != string(tagExpected.Value) {
-			t.Fatalf(`unexpected tag value for key %q at #%d,%d; got %q; want %q`, tag.Key, pos, i, tag.Value, tagExpected.Value)
+			t.Fatalf(`unexpected tag value for key %q at #%d,%d; got %q; want %q; metricGot=%s, metricExpected=%s`,
+				tag.Key, pos, i, tag.Value, tagExpected.Value, mn.String(), mnExpected.String())
 		}
 	}
 }

app/vmselect/promql/transform.go

@@ -1823,8 +1823,15 @@ func newTransformFuncSort(isDesc bool) transformFunc {
 			b := rvs[j].Values
 			n := len(a) - 1
 			for n >= 0 {
-				if !math.IsNaN(a[n]) && !math.IsNaN(b[n]) && a[n] != b[n] {
-					break
+				if !math.IsNaN(a[n]) {
+					if math.IsNaN(b[n]) {
+						return false
+					}
+					if a[n] != b[n] {
+						break
+					}
+				} else if !math.IsNaN(b[n]) {
+					return true
 				}
 				n--
 			}

docs/CHANGELOG.md

@@ -8,6 +8,8 @@ sort: 15
 
 * FEATURE: add `-search.maxSamplesPerSeries` command-line flag for limiting the number of raw samples a single query could process per each time series. This option can prevent from out of memory errors when a query processes tens of millions of raw samples per series. See [this issue](https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1067).
 
+* BUGFIX: return series with `a op b` labels and `N` values for `(a op b) default N` if `(a op b)` returns series with all NaN values. Previously such series were removed.
+
 
 ## [v1.63.0](https://github.com/VictoriaMetrics/VictoriaMetrics/releases/tag/v1.63.0)