Merge pull request #1814 from kuzudb/api-data-type-fix

Add new data type support for C, Java, Python API

src/c_api/value.cpp

@@ -164,11 +164,17 @@ kuzu_value* kuzu_value_get_struct_field_value(kuzu_value* value, uint64_t index)
 }
 
 kuzu_value* kuzu_value_get_recursive_rel_node_list(kuzu_value* value) {
-    return kuzu_value_get_list_element(value, 0);
+    auto* c_value = (kuzu_value*)malloc(sizeof(kuzu_value));
+    c_value->_is_owned_by_cpp = true;
+    c_value->_value = RecursiveRelVal::getNodes(static_cast<Value*>(value->_value));
+    return c_value;
 }
 
 kuzu_value* kuzu_value_get_recursive_rel_rel_list(kuzu_value* value) {
-    return kuzu_value_get_list_element(value, 1);
+    auto* c_value = (kuzu_value*)malloc(sizeof(kuzu_value));
+    c_value->_is_owned_by_cpp = true;
+    c_value->_value = RecursiveRelVal::getRels(static_cast<Value*>(value->_value));
+    return c_value;
 }
 
 kuzu_logical_type* kuzu_value_get_data_type(kuzu_value* value) {

src/common/types/value.cpp

@@ -566,7 +566,7 @@ std::string NodeVal::toString(const Value* val) {
 void NodeVal::throwIfNotNode(const Value* val) {
     if (val->getDataType().getLogicalTypeID() != LogicalTypeID::NODE) {
         auto actualType = LogicalTypeUtils::dataTypeToString(val->getDataType().getLogicalTypeID());
-        throw Exception(fmt::format("Expected node type, but got {} type", actualType));
+        throw Exception(fmt::format("Expected NODE type, but got {} type", actualType));
     }
 }
 
@@ -660,7 +660,24 @@ std::unique_ptr<Value> RelVal::copy(const Value* val) {
 void RelVal::throwIfNotRel(const Value* val) {
     if (val->getDataType().getLogicalTypeID() != LogicalTypeID::REL) {
         auto actualType = LogicalTypeUtils::dataTypeToString(val->getDataType().getLogicalTypeID());
-        throw Exception(fmt::format("Expected relationship type, but got {} type", actualType));
+        throw Exception(fmt::format("Expected REL type, but got {} type", actualType));
+    }
+}
+
+Value* RecursiveRelVal::getNodes(const Value* val) {
+    throwIfNotRecursiveRel(val);
+    return val->getListValReference()[0].get();
+}
+
+Value* RecursiveRelVal::getRels(const Value* val) {
+    throwIfNotRecursiveRel(val);
+    return val->getListValReference()[1].get();
+}
+
+void RecursiveRelVal::throwIfNotRecursiveRel(const Value* val) {
+    if (val->getDataType().getLogicalTypeID() != LogicalTypeID::RECURSIVE_REL) {
+        auto actualType = LogicalTypeUtils::dataTypeToString(val->getDataType().getLogicalTypeID());
+        throw Exception(fmt::format("Expected RECURSIVE_REL type, but got {} type", actualType));
     }
 }
 

src/include/c_api/kuzu.h

@@ -101,6 +101,11 @@ KUZU_C_API typedef enum {
     KUZU_ANY = 0,
     KUZU_NODE = 10,
     KUZU_REL = 11,
+    KUZU_RECURSIVE_REL = 12,
+    // SERIAL is a special data type that is used to represent a sequence of INT64 values that are
+    // incremented by 1 starting from 0.
+    KUZU_SERIAL = 13,
+    // fixed size types
     KUZU_BOOL = 22,
     KUZU_INT64 = 23,
     KUZU_INT32 = 24,
@@ -112,9 +117,14 @@ KUZU_C_API typedef enum {
     KUZU_INTERVAL = 30,
     KUZU_FIXED_LIST = 31,
     KUZU_INTERNAL_ID = 40,
+    KUZU_ARROW_COLUMN = 41,
+    // variable size types
     KUZU_STRING = 50,
+    KUZU_BLOB = 51,
     KUZU_VAR_LIST = 52,
     KUZU_STRUCT = 53,
+    KUZU_MAP = 54,
+    KUZU_UNION = 55,
 } kuzu_data_type_id;
 
 // Database

src/include/common/types/value.h

@@ -339,6 +339,26 @@ class RelVal {
     inline static const uint64_t OFFSET = 4;
 };
 
+/**
+ * @brief RecursiveRelVal represents a path in the graph and stores the corresponding rels and nodes
+ * of that path.
+ */
+class RecursiveRelVal {
+public:
+    /**
+     * @return the list of nodes in the recursive rel as a Value.
+     */
+    KUZU_API static Value* getNodes(const Value* val);
+
+    /**
+     * @return the list of rels in the recursive rel as a Value.
+     */
+    KUZU_API static Value* getRels(const Value* val);
+
+private:
+    static void throwIfNotRecursiveRel(const Value* val);
+};
+
 /**
  * @return boolean value.
  */

tools/java_api/src/jni/kuzu_java.cpp

@@ -109,46 +109,8 @@ internalID_t getInternalID(JNIEnv* env, jobject id) {
 }
 
 std::string dataTypeToString(const LogicalType& dataType) {
-    switch (dataType.getLogicalTypeID()) {
-    case LogicalTypeID::ANY:
-        return "ANY";
-    case LogicalTypeID::NODE:
-        return "NODE";
-    case LogicalTypeID::REL:
-        return "REL";
-    case LogicalTypeID::SERIAL:
-        return "SERIAL";
-    case LogicalTypeID::BOOL:
-        return "BOOL";
-    case LogicalTypeID::INT64:
-        return "INT64";
-    case LogicalTypeID::INT32:
-        return "INT32";
-    case LogicalTypeID::INT16:
-        return "INT16";
-    case LogicalTypeID::DOUBLE:
-        return "DOUBLE";
-    case LogicalTypeID::FLOAT:
-        return "FLOAT";
-    case LogicalTypeID::DATE:
-        return "DATE";
-    case LogicalTypeID::TIMESTAMP:
-        return "TIMESTAMP";
-    case LogicalTypeID::INTERVAL:
-        return "INTERVAL";
-    case LogicalTypeID::FIXED_LIST:
-        return "FIXED_LIST";
-    case LogicalTypeID::INTERNAL_ID:
-        return "INTERNAL_ID";
-    case LogicalTypeID::STRING:
-        return "STRING";
-    case LogicalTypeID::VAR_LIST:
-        return "VAR_LIST";
-    case LogicalTypeID::STRUCT:
-        return "STRUCT";
-    default:
-        throw std::invalid_argument("Unimplemented item");
-    }
+    auto typeId = dataType.getLogicalTypeID();
+    return LogicalTypeUtils::dataTypeToString(typeId);
 }
 
 void javaMapToCPPMap(

tools/java_api/src/main/java/com/kuzudb/KuzuDataTypeID.java

@@ -7,6 +7,8 @@ public enum KuzuDataTypeID {
     ANY(0),
     NODE(10),
     REL(11),
+    RECURSIVE_REL(12),
+    SERIAL(13),
     BOOL(22),
     INT64(23),
     INT32(24),
@@ -18,9 +20,13 @@ public enum KuzuDataTypeID {
     INTERVAL(30),
     FIXED_LIST(31),
     INTERNAL_ID(40),
+    ARROW_COLUMN(41),
     STRING(50),
+    BLOB(51),
     VAR_LIST(52),
-    STRUCT(53);
+    STRUCT(53),
+    MAP(54),
+    UNION(55);
 
     public final int value;
 

tools/python_api/src_py/types.py

@@ -3,19 +3,26 @@
 
 class Type(Enum):
     """The type of a value in the database."""
+    ANY = "ANY"
+    NODE = "NODE"
+    REL = "REL"
+    RECURSIVE_REL = "RECURSIVE_REL"
+    SERIAL = "SERIAL"
     BOOL = "BOOL"
     INT64 = "INT64"
     INT32 = "INT32"
     INT16 = "INT16"
     DOUBLE = "DOUBLE"
     FLOAT = "FLOAT"
-    STRING = "STRING"
     DATE = "DATE"
     TIMESTAMP = "TIMESTAMP"
     INTERVAL = "INTERVAL"
-    LIST = "LIST"
-    NODE = "NODE"
-    REL = "REL"
-    NODE_ID = "NODE_ID"
+    FIXED_LIST = "FIXED_LIST"
+    INTERNAL_ID = "INTERNAL_ID"
+    ARROW_COLUMN = "ARROW_COLUMN"
+    STRING = "STRING"
+    BLOB = "BLOB"
+    VAR_LIST = "VAR_LIST"
     STRUCT = "STRUCT"
-    RECURSIVE_REL = "RECURSIVE_REL"
+    MAP = "MAP"
+    UNION = "UNION"