Implement 'GROUP BY' statement

parsers/models.go

@@ -8,6 +8,7 @@ type SelectStmt struct {
 	Count            int
 	Parameters       map[string]interface{}
 	OrderExpressions []OrderExpression
+	GroupBy          []SelectItem
 }
 
 type Table struct {

parsers/nosql/misc_test.go

@@ -63,6 +63,24 @@ func Test_Parse(t *testing.T) {
 		)
 	})
 
+	t.Run("Should parse SELECT with GROUP BY", func(t *testing.T) {
+		testQueryParse(
+			t,
+			`SELECT c.id, c["pk"] FROM c GROUP BY c.id, c.pk`,
+			parsers.SelectStmt{
+				SelectItems: []parsers.SelectItem{
+					{Path: []string{"c", "id"}},
+					{Path: []string{"c", "pk"}},
+				},
+				Table: parsers.Table{Value: "c"},
+				GroupBy: []parsers.SelectItem{
+					{Path: []string{"c", "id"}},
+					{Path: []string{"c", "pk"}},
+				},
+			},
+		)
+	})
+
 	t.Run("Should parse IN function", func(t *testing.T) {
 		testQueryParse(
 			t,

parsers/nosql/nosql.peg

@@ -6,7 +6,7 @@ import "github.com/pikami/cosmium/parsers"
 func makeSelectStmt(
 	columns, table,
 	whereClause interface{}, distinctClause interface{},
-	count interface{}, orderList interface{},
+	count interface{}, groupByClause interface{}, orderList interface{},
 ) (parsers.SelectStmt, error) {
 	selectStmt := parsers.SelectStmt{
 		SelectItems: columns.([]parsers.SelectItem),
@@ -30,6 +30,10 @@ func makeSelectStmt(
 		selectStmt.OrderExpressions = orderExpressions
 	}
 
+    if groupByClause != nil {
+        selectStmt.GroupBy = groupByClause.([]parsers.SelectItem)
+    }
+
 	return selectStmt, nil
 }
 
@@ -149,8 +153,10 @@ SelectStmt <- Select ws
     topClause:TopClause? ws columns:Selection ws
     From ws table:TableName ws
     whereClause:(ws Where ws condition:Condition { return condition, nil })?
+    groupByClause:(ws GroupBy ws columns:ColumnList { return columns, nil })?
     orderByClause:OrderByClause? {
-    return makeSelectStmt(columns, table, whereClause, distinctClause, topClause, orderByClause)
+    return makeSelectStmt(columns, table, whereClause,
+        distinctClause, topClause, groupByClause, orderByClause)
 }
 
 DistinctClause <- "DISTINCT"i
@@ -285,6 +291,8 @@ And <- "AND"i
 
 Or <- "OR"i
 
+GroupBy <- "GROUP"i ws "BY"i
+
 OrderBy <- "ORDER"i ws "BY"i
 
 ComparisonOperator <- ("=" / "!=" / "<" / "<=" / ">" / ">=") {

query_executors/memory_executor/memory_executor.go

@@ -36,12 +36,20 @@ func Execute(query parsers.SelectStmt, data []RowType) []RowType {
 		ctx.orderBy(query.OrderExpressions, result)
 	}
 
-	// Apply select
+	// Apply group
-	selectedData := make([]RowType, 0)
+	isGroupSelect := query.GroupBy != nil && len(query.GroupBy) > 0
-	for _, row := range result {
+	if isGroupSelect {
-		selectedData = append(selectedData, ctx.selectRow(query.SelectItems, row))
+		result = ctx.groupBy(query, result)
+	}
+
+	// Apply select
+	if !isGroupSelect {
+		selectedData := make([]RowType, 0)
+		for _, row := range result {
+			selectedData = append(selectedData, ctx.selectRow(query.SelectItems, row))
+		}
+		result = selectedData
 	}
-	result = selectedData
 
 	// Apply distinct
 	if query.Distinct {
@@ -182,6 +190,11 @@ func (c memoryExecutorContext) getFieldValue(field parsers.SelectItem, row RowTy
 		return typedValue.Value
 	}
 
+	rowValue := row
+	if array, isArray := row.([]RowType); isArray {
+		rowValue = array[0]
+	}
+
 	if field.Type == parsers.SelectItemTypeFunctionCall {
 		var typedValue parsers.FunctionCall
 		var ok bool
@@ -192,82 +205,83 @@ func (c memoryExecutorContext) getFieldValue(field parsers.SelectItem, row RowTy
 
 		switch typedValue.Type {
 		case parsers.FunctionCallStringEquals:
-			return c.strings_StringEquals(typedValue.Arguments, row)
+			return c.strings_StringEquals(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallContains:
-			return c.strings_Contains(typedValue.Arguments, row)
+			return c.strings_Contains(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallEndsWith:
-			return c.strings_EndsWith(typedValue.Arguments, row)
+			return c.strings_EndsWith(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallStartsWith:
-			return c.strings_StartsWith(typedValue.Arguments, row)
+			return c.strings_StartsWith(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallConcat:
-			return c.strings_Concat(typedValue.Arguments, row)
+			return c.strings_Concat(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIndexOf:
-			return c.strings_IndexOf(typedValue.Arguments, row)
+			return c.strings_IndexOf(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallToString:
-			return c.strings_ToString(typedValue.Arguments, row)
+			return c.strings_ToString(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallUpper:
-			return c.strings_Upper(typedValue.Arguments, row)
+			return c.strings_Upper(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallLower:
-			return c.strings_Lower(typedValue.Arguments, row)
+			return c.strings_Lower(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallLeft:
-			return c.strings_Left(typedValue.Arguments, row)
+			return c.strings_Left(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallLength:
-			return c.strings_Length(typedValue.Arguments, row)
+			return c.strings_Length(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallLTrim:
-			return c.strings_LTrim(typedValue.Arguments, row)
+			return c.strings_LTrim(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallReplace:
-			return c.strings_Replace(typedValue.Arguments, row)
+			return c.strings_Replace(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallReplicate:
-			return c.strings_Replicate(typedValue.Arguments, row)
+			return c.strings_Replicate(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallReverse:
-			return c.strings_Reverse(typedValue.Arguments, row)
+			return c.strings_Reverse(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallRight:
-			return c.strings_Right(typedValue.Arguments, row)
+			return c.strings_Right(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallRTrim:
-			return c.strings_RTrim(typedValue.Arguments, row)
+			return c.strings_RTrim(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallSubstring:
-			return c.strings_Substring(typedValue.Arguments, row)
+			return c.strings_Substring(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallTrim:
-			return c.strings_Trim(typedValue.Arguments, row)
+			return c.strings_Trim(typedValue.Arguments, rowValue)
 
 		case parsers.FunctionCallIsDefined:
-			return c.typeChecking_IsDefined(typedValue.Arguments, row)
+			return c.typeChecking_IsDefined(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsArray:
-			return c.typeChecking_IsArray(typedValue.Arguments, row)
+			return c.typeChecking_IsArray(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsBool:
-			return c.typeChecking_IsBool(typedValue.Arguments, row)
+			return c.typeChecking_IsBool(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsFiniteNumber:
-			return c.typeChecking_IsFiniteNumber(typedValue.Arguments, row)
+			return c.typeChecking_IsFiniteNumber(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsInteger:
-			return c.typeChecking_IsInteger(typedValue.Arguments, row)
+			return c.typeChecking_IsInteger(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsNull:
-			return c.typeChecking_IsNull(typedValue.Arguments, row)
+			return c.typeChecking_IsNull(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsNumber:
-			return c.typeChecking_IsNumber(typedValue.Arguments, row)
+			return c.typeChecking_IsNumber(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsObject:
-			return c.typeChecking_IsObject(typedValue.Arguments, row)
+			return c.typeChecking_IsObject(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsPrimitive:
-			return c.typeChecking_IsPrimitive(typedValue.Arguments, row)
+			return c.typeChecking_IsPrimitive(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallIsString:
-			return c.typeChecking_IsString(typedValue.Arguments, row)
+			return c.typeChecking_IsString(typedValue.Arguments, rowValue)
 
 		case parsers.FunctionCallArrayConcat:
-			return c.array_Concat(typedValue.Arguments, row)
+			return c.array_Concat(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallArrayLength:
-			return c.array_Length(typedValue.Arguments, row)
+			return c.array_Length(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallArraySlice:
-			return c.array_Slice(typedValue.Arguments, row)
+			return c.array_Slice(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallSetIntersect:
-			return c.set_Intersect(typedValue.Arguments, row)
+			return c.set_Intersect(typedValue.Arguments, rowValue)
 		case parsers.FunctionCallSetUnion:
-			return c.set_Union(typedValue.Arguments, row)
+			return c.set_Union(typedValue.Arguments, rowValue)
 
 		case parsers.FunctionCallIn:
-			return c.misc_In(typedValue.Arguments, row)
+			return c.misc_In(typedValue.Arguments, rowValue)
 		}
 	}
 
-	value := row
+	value := rowValue
+
 	if len(field.Path) > 1 {
 		for _, pathSegment := range field.Path[1:] {
 			if nestedValue, ok := value.(map[string]interface{}); ok {
@@ -314,6 +328,47 @@ func (c memoryExecutorContext) orderBy(orderBy []parsers.OrderExpression, data [
 	sort.SliceStable(data, less)
 }
 
+func (c memoryExecutorContext) groupBy(selectStmt parsers.SelectStmt, data []RowType) []RowType {
+	groupedRows := make(map[string][]RowType)
+	groupedKeys := make([]string, 0)
+
+	// Group rows by group by columns
+	for _, row := range data {
+		key := c.generateGroupKey(selectStmt.GroupBy, row)
+		if _, ok := groupedRows[key]; !ok {
+			groupedKeys = append(groupedKeys, key)
+		}
+		groupedRows[key] = append(groupedRows[key], row)
+	}
+
+	// Aggregate each group
+	aggregatedRows := make([]RowType, 0)
+	for _, key := range groupedKeys {
+		groupRows := groupedRows[key]
+		aggregatedRow := c.aggregateGroup(selectStmt, groupRows)
+		aggregatedRows = append(aggregatedRows, aggregatedRow)
+	}
+
+	return aggregatedRows
+}
+
+func (c memoryExecutorContext) generateGroupKey(groupByFields []parsers.SelectItem, row RowType) string {
+	var keyBuilder strings.Builder
+	for _, column := range groupByFields {
+		fieldValue := c.getFieldValue(column, row)
+		keyBuilder.WriteString(fmt.Sprintf("%v", fieldValue))
+		keyBuilder.WriteString(":")
+	}
+
+	return keyBuilder.String()
+}
+
+func (c memoryExecutorContext) aggregateGroup(selectStmt parsers.SelectStmt, groupRows []RowType) RowType {
+	aggregatedRow := c.selectRow(selectStmt.SelectItems, groupRows)
+
+	return aggregatedRow
+}
+
 func compareValues(val1, val2 interface{}) int {
 	if reflect.TypeOf(val1) != reflect.TypeOf(val2) {
 		return 1

query_executors/memory_executor/misc_test.go

@@ -59,6 +59,26 @@ func Test_Execute(t *testing.T) {
 		)
 	})
 
+	t.Run("Should execute SELECT with GROUP BY", func(t *testing.T) {
+		testQueryExecute(
+			t,
+			parsers.SelectStmt{
+				SelectItems: []parsers.SelectItem{
+					{Path: []string{"c", "pk"}},
+				},
+				Table: parsers.Table{Value: "c"},
+				GroupBy: []parsers.SelectItem{
+					{Path: []string{"c", "pk"}},
+				},
+			},
+			mockData,
+			[]memoryexecutor.RowType{
+				map[string]interface{}{"pk": 123},
+				map[string]interface{}{"pk": 456},
+			},
+		)
+	})
+
 	t.Run("Should execute IN function", func(t *testing.T) {
 		testQueryExecute(
 			t,