Rework RelID from global to local

CMakeLists.txt

@@ -1,6 +1,6 @@
 cmake_minimum_required(VERSION 3.11)
 
-project(Kuzu VERSION 0.0.1.1 LANGUAGES CXX)
+project(Kuzu VERSION 0.0.1.2 LANGUAGES CXX)
 
 find_package(Threads REQUIRED)
 

src/binder/bind/bind_projection_clause.cpp

@@ -26,8 +26,6 @@ unique_ptr<BoundReturnClause> Binder::bindReturnClause(const ReturnClause& retur
     auto projectionBody = returnClause.getProjectionBody();
     auto boundProjectionExpressions = bindProjectionExpressions(
         projectionBody->getProjectionExpressions(), projectionBody->containsStar());
-    validateProjectionColumnHasNoInternalType(boundProjectionExpressions);
-    // expand node/rel to all of its properties.
     auto statementResult = make_unique<BoundStatementResult>();
     for (auto& expression : boundProjectionExpressions) {
         auto dataType = expression->getDataType();

src/binder/bind_expression/bind_function_expression.cpp

@@ -128,7 +128,7 @@ unique_ptr<Expression> ExpressionBinder::createInternalNodeIDExpression(
         propertyIDPerTable.insert({tableID, INVALID_PROPERTY_ID});
     }
     auto result = make_unique<PropertyExpression>(
-        DataType(NODE_ID), INTERNAL_ID_SUFFIX, node, std::move(propertyIDPerTable));
+        DataType(INTERNAL_ID), INTERNAL_ID_SUFFIX, node, std::move(propertyIDPerTable));
     return result;
 }
 

src/binder/binder.cpp

@@ -88,17 +88,6 @@ void Binder::validateProjectionColumnNamesAreUnique(const expression_vector& exp
     }
 }
 
-void Binder::validateProjectionColumnHasNoInternalType(const expression_vector& expressions) {
-    auto internalTypes = unordered_set<DataTypeID>{NODE_ID};
-    for (auto& expression : expressions) {
-        if (internalTypes.contains(expression->dataType.typeID)) {
-            throw BinderException("Cannot return expression " + expression->getRawName() +
-                                  " with internal type " +
-                                  Types::dataTypeToString(expression->dataType));
-        }
-    }
-}
-
 void Binder::validateProjectionColumnsInWithClauseAreAliased(const expression_vector& expressions) {
     for (auto& expression : expressions) {
         if (!expression->hasAlias()) {

src/catalog/catalog.cpp

@@ -225,7 +225,7 @@ table_id_t CatalogContent::addRelTableSchema(string tableName, RelMultiplicity r
     }
     vector<Property> properties;
     auto propertyID = 0;
-    auto propertyNameDataType = PropertyNameDataType(INTERNAL_ID_SUFFIX, INT64);
+    auto propertyNameDataType = PropertyNameDataType(INTERNAL_ID_SUFFIX, INTERNAL_ID);
     properties.push_back(
         Property::constructRelProperty(propertyNameDataType, propertyID++, tableID));
     for (auto& propertyDefinition : propertyDefinitions) {

src/common/types/types.cpp

@@ -54,8 +54,8 @@ DataType Types::dataTypeFromString(const string& dataTypeString) {
 }
 
 DataTypeID Types::dataTypeIDFromString(const std::string& dataTypeIDString) {
-    if ("NODE_ID" == dataTypeIDString) {
-        return NODE_ID;
+    if ("INTERNAL_ID" == dataTypeIDString) {
+        return INTERNAL_ID;
     } else if ("INT64" == dataTypeIDString) {
         return INT64;
     } else if ("DOUBLE" == dataTypeIDString) {
@@ -93,8 +93,8 @@ string Types::dataTypeToString(DataTypeID dataTypeID) {
         return "NODE";
     case REL:
         return "REL";
-    case NODE_ID:
-        return "NODE_ID";
+    case INTERNAL_ID:
+        return "INTERNAL_ID";
     case BOOL:
         return "BOOL";
     case INT64:
@@ -138,8 +138,8 @@ string Types::dataTypesToString(const vector<DataTypeID>& dataTypeIDs) {
 
 uint32_t Types::getDataTypeSize(DataTypeID dataTypeID) {
     switch (dataTypeID) {
-    case NODE_ID:
-        return sizeof(nodeID_t);
+    case INTERNAL_ID:
+        return sizeof(internalID_t);
     case BOOL:
         return sizeof(uint8_t);
     case INT64:

src/common/types/value.cpp

@@ -27,7 +27,7 @@ Value Value::createDefaultValue(const DataType& dataType) {
         return Value(timestamp_t());
     case INTERVAL:
         return Value(interval_t());
-    case NODE_ID:
+    case INTERNAL_ID:
         return Value(nodeID_t());
     case STRING:
         return Value(string(""));
@@ -67,8 +67,8 @@ Value::Value(kuzu::common::interval_t val_) : dataType{INTERVAL}, isNull_{false}
     val.intervalVal = val_;
 }
 
-Value::Value(kuzu::common::nodeID_t val_) : dataType{NODE_ID}, isNull_{false} {
-    val.nodeIDVal = val_;
+Value::Value(kuzu::common::internalID_t val_) : dataType{INTERNAL_ID}, isNull_{false} {
+    val.internalIDVal = val_;
 }
 
 Value::Value(const char* val_) : dataType{STRING}, isNull_{false} {
@@ -121,8 +121,8 @@ void Value::copyValueFrom(const uint8_t* value) {
     case INTERVAL: {
         val.intervalVal = *((interval_t*)value);
     } break;
-    case NODE_ID: {
-        val.nodeIDVal = *((nodeID_t*)value);
+    case INTERNAL_ID: {
+        val.internalIDVal = *((nodeID_t*)value);
     } break;
     case STRING: {
         strVal = ((ku_string_t*)value)->getAsString();
@@ -162,8 +162,8 @@ void Value::copyValueFrom(const Value& other) {
     case INTERVAL: {
         val.intervalVal = other.val.intervalVal;
     } break;
-    case NODE_ID: {
-        val.nodeIDVal = other.val.nodeIDVal;
+    case INTERNAL_ID: {
+        val.internalIDVal = other.val.internalIDVal;
     } break;
     case STRING: {
         strVal = other.strVal;
@@ -202,8 +202,8 @@ string Value::toString() const {
         return TypeUtils::toString(val.timestampVal);
     case INTERVAL:
         return TypeUtils::toString(val.intervalVal);
-    case NODE_ID:
-        return TypeUtils::toString(val.nodeIDVal);
+    case INTERNAL_ID:
+        return TypeUtils::toString(val.internalIDVal);
     case STRING:
         return strVal;
     case LIST: {

src/function/aggregate_function.cpp

@@ -96,7 +96,7 @@ unique_ptr<AggregateFunction> AggregateFunctionUtil::getMinMaxFunction(
             MinMaxFunction<ku_string_t>::updatePos<FUNC>,
             MinMaxFunction<ku_string_t>::combine<FUNC>, MinMaxFunction<ku_string_t>::finalize,
             inputType, isDistinct);
-    case NODE_ID:
+    case INTERNAL_ID:
         return make_unique<AggregateFunction>(MinMaxFunction<nodeID_t>::initialize,
             MinMaxFunction<nodeID_t>::updateAll<FUNC>, MinMaxFunction<nodeID_t>::updatePos<FUNC>,
             MinMaxFunction<nodeID_t>::combine<FUNC>, MinMaxFunction<nodeID_t>::finalize, inputType,

src/function/built_in_vector_operations.cpp

@@ -326,10 +326,10 @@ void BuiltInVectorOperations::registerListOperations() {
 void BuiltInVectorOperations::registerInternalIDOperation() {
     vector<unique_ptr<VectorOperationDefinition>> definitions;
     definitions.push_back(make_unique<VectorOperationDefinition>(
-        ID_FUNC_NAME, vector<DataTypeID>{NODE}, NODE_ID, nullptr));
+        ID_FUNC_NAME, vector<DataTypeID>{NODE}, INTERNAL_ID, nullptr));
     definitions.push_back(make_unique<VectorOperationDefinition>(
-        ID_FUNC_NAME, vector<DataTypeID>{REL}, INT64, nullptr));
-    vectorOperations.insert({ID_FUNC_NAME, move(definitions)});
+        ID_FUNC_NAME, vector<DataTypeID>{REL}, INTERNAL_ID, nullptr));
+    vectorOperations.insert({ID_FUNC_NAME, std::move(definitions)});
 }
 
 } // namespace function

src/function/vector_hash_operations.cpp

@@ -10,8 +10,8 @@ void VectorHashOperations::computeHash(ValueVector* operand, ValueVector* result
     result->state = operand->state;
     assert(result->dataType.typeID == INT64);
     switch (operand->dataType.typeID) {
-    case NODE_ID: {
-        UnaryHashOperationExecutor::execute<nodeID_t, hash_t>(*operand, *result);
+    case INTERNAL_ID: {
+        UnaryHashOperationExecutor::execute<internalID_t, hash_t>(*operand, *result);
     } break;
     case BOOL: {
         UnaryHashOperationExecutor::execute<bool, hash_t>(*operand, *result);

src/include/binder/binder.h

@@ -142,9 +142,6 @@ class Binder {
     // E.g. ... RETURN a, b AS a
     static void validateProjectionColumnNamesAreUnique(const expression_vector& expressions);
 
-    // E.g. ... RETURN ID(a)
-    static void validateProjectionColumnHasNoInternalType(const expression_vector& expressions);
-
     // E.g. ... WITH COUNT(*) MATCH ...
     static void validateProjectionColumnsInWithClauseAreAliased(
         const expression_vector& expressions);

src/include/common/types/node_id_t.h → src/include/common/types/internal_id_t.h

@@ -5,37 +5,40 @@
 namespace kuzu {
 namespace common {
 
+struct internalID_t;
+typedef internalID_t nodeID_t;
+typedef internalID_t relID_t;
+
 typedef uint64_t table_id_t;
-typedef uint64_t node_offset_t;
+typedef uint64_t offset_t;
 constexpr table_id_t INVALID_TABLE_ID = UINT64_MAX;
-constexpr node_offset_t INVALID_NODE_OFFSET = UINT64_MAX;
+constexpr offset_t INVALID_NODE_OFFSET = UINT64_MAX;
 
-// System representation for nodeID.
-struct nodeID_t {
-    node_offset_t offset;
+// System representation for internalID.
+struct internalID_t {
+    offset_t offset;
     table_id_t tableID;
 
-    nodeID_t() = default;
-    explicit inline nodeID_t(node_offset_t _offset, table_id_t tableID)
-        : offset(_offset), tableID(tableID) {}
+    internalID_t() = default;
+    internalID_t(offset_t offset, table_id_t tableID) : offset(offset), tableID(tableID) {}
 
     // comparison operators
-    inline bool operator==(const nodeID_t& rhs) const {
+    inline bool operator==(const internalID_t& rhs) const {
         return offset == rhs.offset && tableID == rhs.tableID;
     };
-    inline bool operator!=(const nodeID_t& rhs) const {
+    inline bool operator!=(const internalID_t& rhs) const {
         return offset != rhs.offset || tableID != rhs.tableID;
     };
-    inline bool operator>(const nodeID_t& rhs) const {
+    inline bool operator>(const internalID_t& rhs) const {
         return (tableID > rhs.tableID) || (tableID == rhs.tableID && offset > rhs.offset);
     };
-    inline bool operator>=(const nodeID_t& rhs) const {
+    inline bool operator>=(const internalID_t& rhs) const {
         return (tableID > rhs.tableID) || (tableID == rhs.tableID && offset >= rhs.offset);
     };
-    inline bool operator<(const nodeID_t& rhs) const {
+    inline bool operator<(const internalID_t& rhs) const {
         return (tableID < rhs.tableID) || (tableID == rhs.tableID && offset < rhs.offset);
     };
-    inline bool operator<=(const nodeID_t& rhs) const {
+    inline bool operator<=(const internalID_t& rhs) const {
         return (tableID < rhs.tableID) || (tableID == rhs.tableID && offset <= rhs.offset);
     };
 };

src/include/common/types/types.h

@@ -53,7 +53,7 @@ enum DataTypeID : uint8_t {
     TIMESTAMP = 26,
     INTERVAL = 27,
 
-    NODE_ID = 40,
+    INTERNAL_ID = 40,
 
     // variable size types
     STRING = 50,
@@ -80,7 +80,7 @@ class DataType {
     }
     static inline std::vector<DataTypeID> getAllValidTypeIDs() {
         return std::vector<DataTypeID>{
-            NODE_ID, BOOL, INT64, DOUBLE, STRING, DATE, TIMESTAMP, INTERVAL, LIST};
+            INTERNAL_ID, BOOL, INT64, DOUBLE, STRING, DATE, TIMESTAMP, INTERVAL, LIST};
     }
 
     DataType& operator=(const DataType& other);

src/include/common/types/types_include.h

@@ -2,9 +2,9 @@
 
 #include "date_t.h"
 #include "dtime_t.h"
+#include "internal_id_t.h"
 #include "interval_t.h"
 #include "ku_list.h"
 #include "ku_string.h"
-#include "node_id_t.h"
 #include "timestamp_t.h"
 #include "types.h"

src/include/common/types/value.h

@@ -26,7 +26,7 @@ class Value {
     explicit Value(date_t val_);
     explicit Value(timestamp_t val_);
     explicit Value(interval_t val_);
-    explicit Value(nodeID_t val_);
+    explicit Value(internalID_t val_);
     explicit Value(const char* val_);
     explicit Value(const string& val_);
     explicit Value(DataType dataType, vector<unique_ptr<Value>> vals);
@@ -91,7 +91,7 @@ class Value {
         common::date_t dateVal;
         common::timestamp_t timestampVal;
         common::interval_t intervalVal;
-        common::nodeID_t nodeIDVal;
+        common::internalID_t internalIDVal;
     } val;
     std::string strVal;
     vector<unique_ptr<Value>> listVal;
@@ -191,8 +191,8 @@ inline interval_t Value::getValue() const {
 
 template<>
 inline nodeID_t Value::getValue() const {
-    validateType(NODE_ID);
-    return val.nodeIDVal;
+    validateType(INTERNAL_ID);
+    return val.internalIDVal;
 }
 
 template<>
@@ -251,8 +251,8 @@ inline interval_t& Value::getValueReference() {
 
 template<>
 inline nodeID_t& Value::getValueReference() {
-    assert(dataType.typeID == NODE_ID);
-    return val.nodeIDVal;
+    assert(dataType.typeID == INTERNAL_ID);
+    return val.internalIDVal;
 }
 
 template<>

src/include/common/vector/value_vector.h

@@ -53,8 +53,8 @@ class ValueVector {
 
     inline uint8_t* getData() const { return valueBuffer.get(); }
 
-    inline node_offset_t readNodeOffset(uint32_t pos) const {
-        assert(dataType.typeID == NODE_ID);
+    inline offset_t readNodeOffset(uint32_t pos) const {
+        assert(dataType.typeID == INTERNAL_ID);
         return getValue<nodeID_t>(pos).offset;
     }
 

src/include/function/comparison/vector_comparison_operations.h

@@ -18,7 +18,8 @@ class VectorComparisonOperations : public VectorOperations {
                 definitions.push_back(getDefinition<FUNC>(name, leftTypeID, rightTypeID));
             }
         }
-        for (auto& typeID : vector<DataTypeID>{BOOL, STRING, NODE_ID, DATE, TIMESTAMP, INTERVAL}) {
+        for (auto& typeID :
+            vector<DataTypeID>{BOOL, STRING, INTERNAL_ID, DATE, TIMESTAMP, INTERVAL}) {
             definitions.push_back(getDefinition<FUNC>(name, typeID, typeID));
         }
         definitions.push_back(getDefinition<FUNC>(name, DATE, TIMESTAMP));
@@ -75,8 +76,8 @@ class VectorComparisonOperations : public VectorOperations {
             assert(rightTypeID == STRING);
             return BinaryExecFunction<ku_string_t, ku_string_t, uint8_t, FUNC>;
         }
-        case NODE_ID: {
-            assert(rightTypeID == NODE_ID);
+        case INTERNAL_ID: {
+            assert(rightTypeID == INTERNAL_ID);
             return BinaryExecFunction<nodeID_t, nodeID_t, uint8_t, FUNC>;
         }
         case DATE: {
@@ -157,8 +158,8 @@ class VectorComparisonOperations : public VectorOperations {
             assert(rightTypeID == STRING);
             return BinarySelectFunction<ku_string_t, ku_string_t, FUNC>;
         }
-        case NODE_ID: {
-            assert(rightTypeID == NODE_ID);
+        case INTERNAL_ID: {
+            assert(rightTypeID == INTERNAL_ID);
             return BinarySelectFunction<nodeID_t, nodeID_t, FUNC>;
         }
         case DATE: {

src/include/function/hash/hash_operations.h

@@ -45,6 +45,11 @@ struct CombineHash {
     }
 };
 
+template<>
+inline void Hash::operation(const internalID_t& key, hash_t& result) {
+    result = murmurhash64(key.offset) ^ murmurhash64(key.tableID);
+}
+
 template<>
 inline void Hash::operation(const bool& key, hash_t& result) {
     result = murmurhash64(key);
@@ -96,11 +101,6 @@ inline void Hash::operation(const interval_t& key, hash_t& result) {
         combineHashScalar(murmurhash64(key.days), murmurhash64(key.micros)));
 }
 
-template<>
-inline void Hash::operation(const nodeID_t& key, hash_t& result) {
-    result = murmurhash64(key.offset) ^ murmurhash64(key.tableID);
-}
-
 template<>
 inline void Hash::operation(const unordered_set<string>& key, hash_t& result) {
     for (auto&& s : key) {