constants.h

Go to the documentation of this file.

#pragma once
 
#include <array>
#include <cstdint>
#include <string_view>
 
#include "array_utils.h"
#include "types.h"
 
namespace kuzu {
namespace common {
 
extern const char* KUZU_VERSION;
 
#define DEFAULT_VECTOR_CAPACITY_LOG_2 11
#ifndef KUZU_VECTOR_CAPACITY_LOG2
#define VECTOR_CAPACITY_LOG_2 DEFAULT_VECTOR_CAPACITY_LOG_2
#else
#define VECTOR_CAPACITY_LOG_2 KUZU_VECTOR_CAPACITY_LOG2
#endif
#if VECTOR_CAPACITY_LOG_2 > 12
#error "Vector capacity log2 should be less than or equal to 12"
#endif
constexpr uint64_t DEFAULT_VECTOR_CAPACITY = static_cast<uint64_t>(1) << VECTOR_CAPACITY_LOG_2;
 
constexpr double DEFAULT_HT_LOAD_FACTOR = 1.5;
 
// This is the default thread sleep time we use when a thread,
// e.g., a worker thread is in TaskScheduler, needs to block.
constexpr uint64_t THREAD_SLEEP_TIME_WHEN_WAITING_IN_MICROS = 500;
 
constexpr uint64_t DEFAULT_CHECKPOINT_WAIT_TIMEOUT_IN_MICROS = 5000000;
 
// Note that some places use std::bit_ceil to calculate resizes,
// which won't work for values other than 2. If this is changed, those will need to be updated
constexpr uint64_t CHUNK_RESIZE_RATIO = 2;
 
struct InternalKeyword {
    static constexpr char ANONYMOUS[] = "";
    static constexpr char ID[] = "_ID";
    static constexpr char LABEL[] = "_LABEL";
    static constexpr char SRC[] = "_SRC";
    static constexpr char DST[] = "_DST";
    static constexpr char DIRECTION[] = "_DIRECTION";
    static constexpr char LENGTH[] = "_LENGTH";
    static constexpr char NODES[] = "_NODES";
    static constexpr char RELS[] = "_RELS";
    static constexpr char STAR[] = "*";
    static constexpr char PLACE_HOLDER[] = "_PLACE_HOLDER";
    static constexpr char MAP_KEY[] = "KEY";
    static constexpr char MAP_VALUE[] = "VALUE";
 
    static constexpr std::string_view ROW_OFFSET = "_row_offset";
    static constexpr std::string_view SRC_OFFSET = "_src_offset";
    static constexpr std::string_view DST_OFFSET = "_dst_offset";
};
 
enum PageSizeClass : uint8_t {
    REGULAR_PAGE = 0,
    TEMP_PAGE = 1,
};
 
// Currently the system supports files with 2 different pages size, which we refer to as
// PAGE_SIZE and TEMP_PAGE_SIZE. PAGE_SIZE is the default size of the page which is the
// unit of read/write to the database files.
#ifdef KUZU_PAGE_SIZE_LOG2
static constexpr uint64_t PAGE_SIZE_LOG2 = KUZU_PAGE_SIZE_LOG2;
#else
static constexpr uint64_t PAGE_SIZE_LOG2 = 12; // Default to 4KB.
#endif
static constexpr uint64_t KUZU_PAGE_SIZE = static_cast<uint64_t>(1) << PAGE_SIZE_LOG2;
// Page size for files with large pages, e.g., temporary files that are used by operators that
// may require large amounts of memory.
static constexpr uint64_t TEMP_PAGE_SIZE_LOG2 = 18;
static const uint64_t TEMP_PAGE_SIZE = static_cast<uint64_t>(1) << TEMP_PAGE_SIZE_LOG2;
 
struct BufferPoolConstants {
    // If a user does not specify a max size for BM, we by default set the max size of BM to
    // maxPhyMemSize * DEFAULT_PHY_MEM_SIZE_RATIO_FOR_BM.
    static constexpr double DEFAULT_PHY_MEM_SIZE_RATIO_FOR_BM = 0.8;
// The default max size for a VMRegion.
#ifdef __32BIT__
    static constexpr uint64_t DEFAULT_VM_REGION_MAX_SIZE = (uint64_t)1 << 30; // (1GB)
#else
    static constexpr uint64_t DEFAULT_VM_REGION_MAX_SIZE = static_cast<uint64_t>(1) << 43; // (8TB)
#endif
    static constexpr uint64_t DEFAULT_BUFFER_POOL_SIZE_FOR_TESTING = 1ull << 28; // (256MB)
};
 
struct StorageConstants {
    static constexpr char OVERFLOW_FILE_SUFFIX[] = ".ovf";
    static constexpr char WAL_FILE_SUFFIX[] = ".wal";
    static constexpr char SHADOWING_SUFFIX[] = ".shadow";
    static constexpr char INDEX_FILE_SUFFIX[] = ".hindex";
    static constexpr char CATALOG_FILE_NAME[] = "catalog.kz";
    static constexpr char CATALOG_FILE_NAME_FOR_WAL[] = "catalog.shadow";
    static constexpr char DATA_FILE_NAME[] = "data.kz";
    static constexpr char METADATA_FILE_NAME[] = "metadata.kz";
    static constexpr char METADATA_FILE_NAME_FOR_WAL[] = "metadata.shadow";
    static constexpr char LOCK_FILE_NAME[] = ".lock";
 
    // The number of pages that we add at one time when we need to grow a file.
    static constexpr uint64_t PAGE_GROUP_SIZE_LOG2 = 10;
    static constexpr uint64_t PAGE_GROUP_SIZE = static_cast<uint64_t>(1) << PAGE_GROUP_SIZE_LOG2;
    static constexpr uint64_t PAGE_IDX_IN_GROUP_MASK =
        (static_cast<uint64_t>(1) << PAGE_GROUP_SIZE_LOG2) - 1;
 
    static constexpr uint64_t NODE_GROUP_SIZE_LOG2 = 17; // 64 * 2048 nodes per group
    static constexpr uint64_t NODE_GROUP_SIZE = static_cast<uint64_t>(1) << NODE_GROUP_SIZE_LOG2;
 
    static constexpr double PACKED_CSR_DENSITY = 0.8;
    static constexpr double LEAF_HIGH_CSR_DENSITY = 1.0;
    // The number of CSR lists in a leaf region.
    static constexpr uint64_t CSR_LEAF_REGION_SIZE_LOG2 = 10;
    static constexpr uint64_t CSR_LEAF_REGION_SIZE = static_cast<uint64_t>(1)
                                                     << CSR_LEAF_REGION_SIZE_LOG2;
 
    static constexpr uint64_t MAX_NUM_ROWS_IN_TABLE = static_cast<uint64_t>(1) << 62;
 
    static constexpr char TEMP_SPILLING_FILE_NAME[] = ".tmp";
};
 
// Hash Index Configurations
struct HashIndexConstants {
    static constexpr uint16_t SLOT_CAPACITY_BYTES = 256;
};
 
struct CopyConstants {
    // Initial size of buffer for CSV Reader.
    static constexpr uint64_t INITIAL_BUFFER_SIZE = 16384;
    // This means that we will usually read the entirety of the contents of the file we need for a
    // block in one read request. It is also very small, which means we can parallelize small files
    // efficiently.
    static constexpr uint64_t PARALLEL_BLOCK_SIZE = INITIAL_BUFFER_SIZE / 2;
 
    static constexpr const char* IGNORE_ERRORS_OPTION_NAME = "IGNORE_ERRORS";
    static constexpr const char* BOOL_CSV_PARSING_OPTIONS[] = {"HEADER", "PARALLEL",
        "LIST_UNBRACED", "AUTODETECT", "AUTO_DETECT", CopyConstants::IGNORE_ERRORS_OPTION_NAME};
    static constexpr bool DEFAULT_CSV_HAS_HEADER = false;
    static constexpr bool DEFAULT_CSV_PARALLEL = true;
 
    // Default configuration for csv file parsing
    static constexpr const char* STRING_CSV_PARSING_OPTIONS[] = {"ESCAPE", "DELIM", "DELIMITER",
        "QUOTE"};
    static constexpr char DEFAULT_CSV_ESCAPE_CHAR = '"';
    static constexpr char DEFAULT_CSV_DELIMITER = ',';
    static constexpr bool DEFAULT_CSV_ALLOW_UNBRACED_LIST = false;
    static constexpr char DEFAULT_CSV_QUOTE_CHAR = '"';
    static constexpr char DEFAULT_CSV_LIST_BEGIN_CHAR = '[';
    static constexpr char DEFAULT_CSV_LIST_END_CHAR = ']';
    static constexpr bool DEFAULT_IGNORE_ERRORS = false;
    static constexpr bool DEFAULT_CSV_AUTO_DETECT = true;
    static constexpr bool DEFAULT_CSV_SET_DIALECT = false;
    static constexpr std::array DEFAULT_CSV_DELIMITER_SEARCH_SPACE = {',', ';', '\t', '|'};
    static constexpr std::array DEFAULT_CSV_QUOTE_SEARCH_SPACE = {'"', '\''};
    static constexpr std::array DEFAULT_CSV_ESCAPE_SEARCH_SPACE = {'"', '\\', '\''};
    static constexpr uint64_t PANDAS_PARTITION_COUNT = 50 * DEFAULT_VECTOR_CAPACITY;
 
    static constexpr const char* INT_CSV_PARSING_OPTIONS[] = {"SKIP", "SAMPLE_SIZE"};
    static constexpr uint64_t DEFAULT_CSV_SKIP_NUM = 0;
    static constexpr uint64_t DEFAULT_CSV_TYPE_DEDUCTION_SAMPLE_SIZE = 256;
 
    // metadata columns used to populate CSV warnings
    static constexpr std::array SHARED_WARNING_DATA_COLUMN_NAMES = {"blockIdx", "offsetInBlock",
        "startByteOffset", "endByteOffset"};
    static constexpr std::array SHARED_WARNING_DATA_COLUMN_TYPES = {LogicalTypeID::UINT64,
        LogicalTypeID::UINT32, LogicalTypeID::UINT64, LogicalTypeID::UINT64};
    static constexpr column_id_t SHARED_WARNING_DATA_NUM_COLUMNS =
        SHARED_WARNING_DATA_COLUMN_NAMES.size();
 
    static constexpr std::array CSV_SPECIFIC_WARNING_DATA_COLUMN_NAMES = {"fileIdx"};
    static constexpr std::array CSV_SPECIFIC_WARNING_DATA_COLUMN_TYPES = {LogicalTypeID::UINT32};
 
    static constexpr std::array CSV_WARNING_DATA_COLUMN_NAMES =
        arrayConcat(SHARED_WARNING_DATA_COLUMN_NAMES, CSV_SPECIFIC_WARNING_DATA_COLUMN_NAMES);
    static constexpr std::array CSV_WARNING_DATA_COLUMN_TYPES =
        arrayConcat(SHARED_WARNING_DATA_COLUMN_TYPES, CSV_SPECIFIC_WARNING_DATA_COLUMN_TYPES);
    static constexpr column_id_t CSV_WARNING_DATA_NUM_COLUMNS =
        CSV_WARNING_DATA_COLUMN_NAMES.size();
    static_assert(CSV_WARNING_DATA_NUM_COLUMNS == CSV_WARNING_DATA_COLUMN_TYPES.size());
 
    static constexpr column_id_t MAX_NUM_WARNING_DATA_COLUMNS = CSV_WARNING_DATA_NUM_COLUMNS;
};
 
struct PlannerKnobs {
    static constexpr double NON_EQUALITY_PREDICATE_SELECTIVITY = 0.1;
    static constexpr double EQUALITY_PREDICATE_SELECTIVITY = 0.01;
    static constexpr uint64_t BUILD_PENALTY = 2;
    // Avoid doing probe to build SIP if we have to accumulate a probe side that is much bigger than
    // build side. Also avoid doing build to probe SIP if probe side is not much bigger than build.
    static constexpr uint64_t SIP_RATIO = 5;
};
 
struct OrderByConstants {
    static constexpr uint64_t NUM_BYTES_FOR_PAYLOAD_IDX = 8;
    static constexpr uint64_t MIN_SIZE_TO_REDUCE = common::DEFAULT_VECTOR_CAPACITY * 5;
    static constexpr uint64_t MIN_LIMIT_RATIO_TO_REDUCE = 2;
};
 
struct ParquetConstants {
    static constexpr uint64_t PARQUET_DEFINE_VALID = 65535;
    static constexpr const char* PARQUET_MAGIC_WORDS = "PAR1";
    // We limit the uncompressed page size to 100MB.
    // The max size in Parquet is 2GB, but we choose a more conservative limit.
    static constexpr uint64_t MAX_UNCOMPRESSED_PAGE_SIZE = 100000000;
    // Dictionary pages must be below 2GB. Unlike data pages, there's only one dictionary page.
    // For this reason we go with a much higher, but still a conservative upper bound of 1GB.
    static constexpr uint64_t MAX_UNCOMPRESSED_DICT_PAGE_SIZE = 1e9;
    // The maximum size a key entry in an RLE page takes.
    static constexpr uint64_t MAX_DICTIONARY_KEY_SIZE = sizeof(uint32_t);
    // The size of encoding the string length.
    static constexpr uint64_t STRING_LENGTH_SIZE = sizeof(uint32_t);
    static constexpr uint64_t MAX_STRING_STATISTICS_SIZE = 10000;
    static constexpr uint64_t PARQUET_INTERVAL_SIZE = 12;
    static constexpr uint64_t PARQUET_UUID_SIZE = 16;
};
 
struct ExportCSVConstants {
    static constexpr const char* DEFAULT_CSV_NEWLINE = "\n";
    static constexpr const char* DEFAULT_NULL_STR = "";
    static constexpr bool DEFAULT_FORCE_QUOTE = false;
    static constexpr uint64_t DEFAULT_CSV_FLUSH_SIZE = 4096 * 8;
};
 
struct ImportDBConstants {
    static constexpr char SCHEMA_NAME[] = "schema.cypher";
    static constexpr char COPY_NAME[] = "copy.cypher";
};
 
struct WarningConstants {
    static constexpr std::array WARNING_TABLE_COLUMN_NAMES{"query_id", "message", "file_path",
        "line_number", "skipped_line_or_record"};
    static constexpr std::array WARNING_TABLE_COLUMN_DATA_TYPES{LogicalTypeID::UINT64,
        LogicalTypeID::STRING, LogicalTypeID::STRING, LogicalTypeID::UINT64, LogicalTypeID::STRING};
    static constexpr uint64_t WARNING_TABLE_NUM_COLUMNS = WARNING_TABLE_COLUMN_NAMES.size();
 
    static_assert(WARNING_TABLE_COLUMN_DATA_TYPES.size() == WARNING_TABLE_NUM_COLUMNS);
};
 
static constexpr char ATTACHED_KUZU_DB_TYPE[] = "KUZU";
 
static constexpr char LOCAL_DB_NAME[] = "local(kuzu)";
 
constexpr auto DECIMAL_PRECISION_LIMIT = 38;
 
static constexpr char SCAN_JSON_FUNC_NAME[] = "READ_JSON";
 
} // namespace common
} // namespace kuzu