Merge pull request #2622 from influxdb/select-distinct-tag-customer-error-2612

Select distinct tag custom error

CHANGELOG.md

@@ -16,6 +16,7 @@
 - [#2596](https://github.com/influxdb/influxdb/pull/2596): RC30: `panic: runtime error: index out of range` when insert data points.
 - [#2592](https://github.com/influxdb/influxdb/pull/2592): Should return an error if user attempts to group by a field.
 - [#2499](https://github.com/influxdb/influxdb/pull/2499): Issuing a select query with tag as a values causes panic.
+- [#2612](https://github.com/influxdb/influxdb/pull/2612): Query planner should validate distinct is passed a field.
 
 ## PRs
 - [#2569](https://github.com/influxdb/influxdb/pull/2569): Add derivative functions

cmd/influxd/server_integration_test.go

@@ -705,29 +705,59 @@ func runTestsData(t *testing.T, testName string, nodes Cluster, database, retent
 		},
 		{
 			reset: true,
-			name:  "distincts",
+			name:  "distinct as call",
 			write: `{"database" : "%DB%", "retentionPolicy" : "%RP%", "points": [
-				{"name": "cpu", "time": "2000-01-01T00:00:00Z", "fields": {"value": 30}},
-				{"name": "cpu", "time": "2000-01-01T00:00:10Z", "fields": {"value": 20}},
-				{"name": "cpu", "time": "2000-01-01T00:00:20Z", "fields": {"value": 30}},
-				{"name": "cpu", "time": "2000-01-01T00:00:30Z", "fields": {"value": 100}}
+				{"name": "cpu", "time": "2000-01-01T00:00:00Z", "tags": {"host": "server01"}, "fields": {"value": 30}},
+				{"name": "cpu", "time": "2000-01-01T00:00:10Z", "tags": {"host": "server02"}, "fields": {"value": 20}},
+				{"name": "cpu", "time": "2000-01-01T00:00:20Z", "tags": {"host": "server03"}, "fields": {"value": 30}},
+				{"name": "cpu", "time": "2000-01-01T00:00:30Z", "tags": {"host": "server03"}, "fields": {"value": 100}}
 			]}`,
 			query:    `SELECT distinct(value) FROM cpu`,
 			queryDb:  "%DB%",
 			expected: `{"results":[{"series":[{"name":"cpu","columns":["time","distinct"],"values":[["1970-01-01T00:00:00Z",[20,30,100]]]}]}]}`,
 		},
 		{
-			name:     "distincts alt syntax",
+			name:     "distinct alt syntax",
 			query:    `SELECT distinct value FROM cpu`,
 			queryDb:  "%DB%",
 			expected: `{"results":[{"series":[{"name":"cpu","columns":["time","distinct"],"values":[["1970-01-01T00:00:00Z",[20,30,100]]]}]}]}`,
 		},
+		{
+			name:     "distinct select tag",
+			query:    `SELECT distinct(host) FROM cpu`,
+			queryDb:  "%DB%",
+			expected: `{"results":[{"error":"host isn't a field on measurement cpu; to query the unique values for a tag use SHOW TAG VALUES FROM cpu WITH KEY = \"host\""}]}`,
+		},
+		{
+			name:     "distinct alt select tag",
+			query:    `SELECT distinct host FROM cpu`,
+			queryDb:  "%DB%",
+			expected: `{"results":[{"error":"host isn't a field on measurement cpu; to query the unique values for a tag use SHOW TAG VALUES FROM cpu WITH KEY = \"host\""}]}`,
+		},
 		{
 			name:     "count distinct",
 			query:    `SELECT count(distinct value) FROM cpu`,
 			queryDb:  "%DB%",
 			expected: `{"results":[{"series":[{"name":"cpu","columns":["time","count"],"values":[["1970-01-01T00:00:00Z",3]]}]}]}`,
 		},
+		{
+			name:     "count distinct as call",
+			query:    `SELECT count(distinct(value)) FROM cpu`,
+			queryDb:  "%DB%",
+			expected: `{"results":[{"series":[{"name":"cpu","columns":["time","count"],"values":[["1970-01-01T00:00:00Z",3]]}]}]}`,
+		},
+		{
+			name:     "count distinct select tag",
+			query:    `SELECT count(distinct host) FROM cpu`,
+			queryDb:  "%DB%",
+			expected: `{"results":[{"error":"host isn't a field on measurement cpu; count(distinct) on tags isn't yet supported"}]}`,
+		},
+		{
+			name:     "count distinct as call select tag",
+			query:    `SELECT count(distinct(host)) FROM cpu`,
+			queryDb:  "%DB%",
+			expected: `{"results":[{"error":"host isn't a field on measurement cpu; count(distinct) on tags isn't yet supported"}]}`,
+		},
 		{
 			reset: true,
 			name:  "aggregations",

influxql/engine.go

@@ -702,7 +702,7 @@ func (m *MapReduceJob) processRawResults(values []*rawQueryMapOutput) *Row {
 				selectNames[0], selectNames[i] = selectNames[i], selectNames[0]
 			}
 			hasTime = true
-      break
+			break
 		}
 	}
 

tx.go

@@ -374,6 +374,8 @@ func (l *LocalMapper) Begin(c *influxql.Call, startingTime int64, chunkSize int)
 	l.chunkSize = chunkSize
 	l.tmin = startingTime
 
+	var isCountDistinct bool
+
 	// determine if this is a raw data query with a single field, multiple fields, or an aggregate
 	var fieldName string
 	if c == nil { // its a raw data query
@@ -400,17 +402,27 @@ func (l *LocalMapper) Begin(c *influxql.Call, startingTime int64, chunkSize int)
 			if c.Name != "count" {
 				return fmt.Errorf("aggregate call didn't contain a field %s", c.String())
 			}
+			isCountDistinct = true
 			fieldName = lit.Val
 		default:
 			return fmt.Errorf("aggregate call didn't contain a field %s", c.String())
 		}
+
+		isCountDistinct = isCountDistinct || (c.Name == "count" && nested.Name == "distinct")
 	}
 
 	// set up the field info if a specific field was set for this mapper
 	if fieldName != "" {
 		f := l.decoder.FieldByName(fieldName)
 		if f == nil {
-			return fmt.Errorf("%s isn't a field on measurement %s", fieldName, l.job.MeasurementName)
+			switch {
+			case c.Name == "distinct":
+				return fmt.Errorf("%s isn't a field on measurement %s; to query the unique values for a tag use SHOW TAG VALUES FROM %s WITH KEY = %q", fieldName, l.job.MeasurementName, l.job.MeasurementName, fieldName)
+			case isCountDistinct:
+				return fmt.Errorf("%s isn't a field on measurement %s; count(distinct) on tags isn't yet supported", fieldName, l.job.MeasurementName)
+			default:
+				return fmt.Errorf("%s isn't a field on measurement %s", fieldName, l.job.MeasurementName)
+			}
 		}
 		l.fieldID = f.ID
 		l.fieldName = f.Name