Core Apache Iceberg C++ APIs. More...

Namespaces
namespace	arrow
	Arrow-backed IO and data conversion APIs.

namespace	avro
	Avro reader, writer, and metadata APIs.

namespace	hive
	Hive catalog APIs.

namespace	parquet
	Parquet reader, writer, and registration APIs.

namespace	puffin
	Puffin file metadata, reader, and writer APIs.

namespace	rest
	REST catalog APIs.

namespace	retry
	Retry policy APIs.

namespace	sql
	SQL catalog APIs.

namespace	table
	Table update requirement APIs.

namespace	util
	Utility APIs.

Classes
class	AddedRowsScanTask
	A scan task for inserts generated by adding a data file to the table. More...

class	Aggregate
	Base aggregate holding an operation and a term. More...

class	AggregateEvaluator
	Evaluates bound aggregates over StructLike data. More...

struct	AliveRowSelection
	Result of ComputeAliveRows: indices of rows not matched by any delete. More...

class	And
	An Expression that represents a logical AND operation between two expressions. More...

class	ArrayLike
	An immutable array-like wrapper. More...

class	ArrowArrayArrayLike
	Wrapper for one row of a list-typed ArrowArray. More...

class	ArrowArrayMapLike
	Wrapper for one row of a map-typed ArrowArray. More...

class	ArrowArrayStructLike
	Wrapper for one row of a struct-typed ArrowArray. More...

class	AssignFreshIdVisitor
	Assigns fresh IDs to all fields in the schema. More...

class	Base64
	Utilities for base64 and base64url encoding and decoding. More...

class	BaseInheritableMetadata
	Base implementation of InheritableMetadata that handles standard inheritance rules. More...

class	BinaryType
	A data type representing an arbitrary-length byte sequence. More...

class	Binder

struct	BlobMetadata
	Metadata about a statistics or indices blob. More...

class	BooleanType
	A data type representing a boolean (true or false). More...

class	Bound
	Interface for bound expressions that can be evaluated. More...

class	BoundAggregate
	Base class for bound aggregates. More...

class	BoundingBox
	A pair of lower and upper geospatial column bounds. More...

class	BoundLiteralPredicate
	Bound literal predicate (comparison against a single value). More...

class	BoundPredicate
	Bound predicates contain bound terms and can be evaluated. More...

class	BoundReference
	A reference to a bound field. More...

class	BoundSetPredicate
	Bound set predicate (membership testing against a set of values). More...

class	BoundTerm
	Base class for bound terms. More...

class	BoundTransform
	A bound transform expression. More...

class	BoundUnaryPredicate
	Bound unary predicate (null, not-null, etc.). More...

class	BoundVisitor
	Visitor for bound expressions. More...

class	BucketTransform
	Bucket transform that hashes input values into N buckets. More...

class	BucketUtils

class	Catalog
	A Catalog API for table create, drop, and load operations. More...

class	CatalogUtil
	Utilities for working with catalogs. More...

class	CerrLogger
	Logger that writes one line per record to std::cerr. More...

class	ChangelogScanTask
	A scan task for reading changelog entries between snapshots. More...

class	CommitMetrics
	Live commit metrics collected during a table commit operation. More...

struct	CommitMetricsResult
	Immutable snapshot of commit metrics for use in CommitReport. More...

struct	CommitReport
	Report generated after a commit operation. More...

class	CompositeMetricsReporter
	A MetricsReporter that delegates to multiple reporters. More...

class	ConfigBase

struct	ContentFileUtil
	Utility functions for content files. More...

class	Conversions
	Conversion utilities for primitive types. More...

class	CopyInheritableMetadata
	Metadata inheritance for copying manifests before commit. More...

class	CountAggregate
	Base class for COUNT aggregates. More...

class	Counter
	Abstract counter for tracking event totals. More...

struct	CounterResult
	Serialisable snapshot of a single Counter measurement. More...

class	CountNonNullAggregate
	COUNT(term) aggregate. More...

class	CountNullAggregate
	COUNT_NULL(term) aggregate. More...

class	CountStarAggregate
	COUNT(*) aggregate. More...

struct	DataFile
	DataFile carries data file path, partition tuple, metrics, ... More...

class	DataFileSet
	A set of DataFile pointers with insertion order preserved and deduplicated by file path. More...

class	DataFileStructLike
	StructLike wrapper for DataFile metadata. More...

struct	DataOperation
	Data operation that produce snapshots. More...

class	DataTableScan
	A scan that reads data files and applies delete files to filter rows. More...

class	DataWriter
	Writer for Iceberg data files. More...

struct	DataWriterOptions
	Options for creating a DataWriter. More...

class	DateType
	A data type representing a calendar date without reference to a timezone or time. More...

class	DayTransform
	Day transform that extracts the number of days from timestamp inputs since the epoch. More...

class	Decimal
	Represents 128-bit fixed-point decimal numbers. The max decimal precision that can be safely represented is 38 significant digits. More...

class	DecimalType
	A data type representing a fixed-precision decimal. More...

class	DefaultCounter
	Thread-safe counter backed by std::atomic<int64_t>. More...

struct	DefaultError
	/brief Default error trait More...

class	DefaultMetricsContext
	MetricsContext backed by DefaultCounter and DefaultTimer instances. More...

class	DefaultTimer
	Thread-safe timer backed by std::atomic<int64_t>. More...

class	DeleteCounter
	Counts rows removed by delete filters. More...

class	DeletedDataFileScanTask
	A scan task for deletes generated by removing a data file from the table. More...

class	DeleteFileIndex
	An index of delete files by sequence number. More...

class	DeleteFiles
	API for deleting data files from a table. More...

class	DeleteFileSet
	A set of delete-file pointers deduplicated by delete-file identity. More...

class	DeleteFilter
	Concrete batch-oriented delete filter for merge-on-read data batches. More...

class	DeleteLoader
	Loads delete files and constructs in-memory delete indexes. More...

struct	DeleteWriteResult
	File metadata for deletion vectors produced by DVWriter. More...

class	DoubleType
	A data type representing a 64-bit (double precision) IEEE-754 float. More...

class	DVWriter
	A deletion vector file writer. More...

struct	DVWriterOptions
	Options for creating a DVWriter. More...

class	EmptyInheritableMetadata
	Empty implementation that applies no inheritance. More...

struct	EncryptedKey
	Represents an encrypted table key entry. More...

class	EqualityDeleteWriter
	Writer for Iceberg equality delete files. More...

struct	EqualityDeleteWriterOptions
	Options for creating an EqualityDeleteWriter. More...

struct	Error
	Error with a kind and a message. More...

class	ErrorCollector
	Base class for collecting errors in the builder pattern. More...

class	Evaluator
	Evaluates an Expression against data rows. More...

class	Executor
	Schedules iceberg-cpp internal planning tasks. More...

class	ExpireSnapshots
	API for removing old snapshots from a table. More...

class	Expression
	Represents a boolean expression tree. More...

class	ExpressionError
	Exception thrown when expression construction fails. More...

class	Expressions
	Fluent APIs to create expressions. More...

class	ExpressionVisitor
	Base visitor for traversing expression trees. More...

class	False
	An expression that is always false. More...

class	FastAppend
	API for appending new files in a table. More...

struct	FieldMetrics
	Field-level metrics for a single column. More...

struct	FieldProjection
	A field schema partner to carry projection information. More...

class	FileIO
	Pluggable module for reading, writing, and deleting files. More...

class	FileIORegistry
	Registry for FileIO implementations. More...

class	FileScanTask
	Task representing a data file and its corresponding delete files. More...

class	FileScanTaskReader
	Opens a FileScanTask as an ArrowArrayStream. More...

class	FileWriter
	Base interface for data file writers. More...

class	FixedType
	A data type representing a fixed-length bytestring. More...

class	FloatType
	A data type representing a 32-bit (single precision) IEEE-754 float. More...

class	FnOnce

class	FnOnce< R(Args...)>

class	GeographyType
	A data type representing OGC geography in WKB format. More...

class	GeometryType
	A data type representing OGC geometry in WKB format. More...

class	GeospatialBound
	A point used as a lower or upper geospatial column bound. More...

class	GetProjectedIdsVisitor
	Visitor for getting projected field IDs. More...

class	HistoryTable
	History metadata table. More...

class	HourTransform
	Hour transform that extracts the number of hours from timestamp inputs since the epoch. More...

class	IcebergError
	Base exception class for exceptions thrown by the Iceberg library. More...

class	IdentityTransform
	Identity transform that returns the input unchanged. More...

class	IdToFieldVisitor
	Visitor for building a map from field ID to SchemaField reference. More...

class	InclusiveMetricsEvaluator

class	IncrementalAppendScan
	A scan that reads data files added between snapshots (incremental appends). More...

class	IncrementalChangelogScan
	A scan that reads changelog entries between snapshots. More...

class	IncrementalScan
	A base template class for incremental scans that read changes between snapshots, and return scan tasks of the specified type. More...

class	InheritableMetadata
	Interface for applying inheritable metadata to manifest entries. More...

class	InheritableMetadataFactory
	Factory for creating InheritableMetadata instances. More...

class	InMemoryCatalog
	An in-memory implementation of the Iceberg Catalog interface. More...

class	InputFile
	Handle for opening a readable file. More...

class	IntType
	A data type representing a 32-bit signed integer. More...

class	IsBoundVisitor

class	Lazy

class	ListType
	A data type representing a list of values. More...

class	Literal
	Literal is a literal value that is associated with a primitive type. More...

struct	LiteralHash

struct	LiteralTraits

struct	LiteralValueHash
	Hash function for Literal to facilitate use in unordered containers. More...

class	LocationProvider
	Interface for providing data file locations. More...

class	LocationUtil

struct	LogAttribute
	A structured key/value attribute attached to a log record. More...

class	Logger
	Pluggable logging sink. More...

class	Loggers
	Registry of logger factories, mirroring MetricsReporters. More...

class	LogMessage
	A single log record handed to a Logger. More...

class	LongType
	A data type representing a 64-bit signed integer. More...

struct	ManifestEntry
	A manifest is an immutable Avro file that lists data files or delete files, along with each file's partition data tuple, metrics, and tracking information. More...

class	ManifestEvaluator
	Evaluates an Expression against manifest. More...

struct	ManifestFile
	Entry in a manifest list. More...

class	ManifestFileStructLike
	StructLike wrapper for ManifestFile. More...

class	ManifestFilterManager
	Filters an existing snapshot's manifest list. More...

class	ManifestGroup
	Coordinates reading manifest files and producing scan tasks. More...

struct	ManifestList

class	ManifestListReader
	Read manifest files from a manifest list file. More...

class	ManifestListWriter
	Write manifest files to a manifest list file. More...

class	ManifestMergeManager
	Merges small manifests into larger ones using greedy bin-packing. More...

class	ManifestReader
	Read manifest entries from a manifest file. More...

class	ManifestWriter
	Write manifest entries to a manifest file. More...

class	MapLike
	An immutable map-like wrapper. More...

struct	MappedField
	An immutable mapping between a field ID and a set of names. More...

class	MappedFields
	A list of field mappings for child field of structs, maps, and lists. More...

class	MapType
	A data type representing a dictionary of values. More...

class	MaxAggregate
	Bound MAX aggregate. More...

class	MergeAppend
	API for appending new files in a table while merging manifests. More...

class	MergingSnapshotUpdate
	Abstract base class for merge-based snapshot write operations. More...

struct	MetadataColumns
	A class containing constants and utility methods for metadata columns. More...

struct	MetadataLogEntry
	Represents a metadata log entry. More...

class	MetadataTable
	Base class for Iceberg metadata tables. More...

struct	Metrics
	Iceberg file format metrics. More...

class	MetricsConfig
	Configuration for collecting column metrics for an Iceberg table. More...

class	MetricsContext
	Factory for creating named Counter and Timer instances. More...

struct	MetricsMode

class	MetricsReporter
	Interface for reporting metrics from Iceberg operations. More...

class	MetricsReporters
	Factory class for creating and managing MetricsReporter instances. More...

class	MinAggregate
	Bound MIN aggregate. More...

class	MonthTransform
	Month transform that extracts the number of months from timestamp inputs since the epoch. More...

class	NamedReference
	A reference to an unbound named field. More...

class	NameMapping
	Represents a mapping from external schema names to Iceberg type IDs. More...

struct	Namespace
	A namespace in a catalog. More...

class	NameToIdVisitor
	Visitor for building maps from field name to field ID and field ID to field name. More...

class	NestedType
	A data type that has child fields. More...

class	Not
	An Expression that represents logical NOT operation. More...

class	Or
	An Expression that represents a logical OR operation between two expressions. More...

class	OutputFile
	Handle for creating a writable file. More...

class	OverwriteFiles
	API for overwriting files in a table. More...

class	PartitionField
	a field with its transform. More...

struct	PartitionFieldSummary
	Field summary for partition field in the spec. More...

class	PartitionFieldSummaryArrayLike
	ArrayLike wrapper for a vector of PartitionFieldSummary. More...

class	PartitionFieldSummaryStructLike
	StructLike wrapper for PartitionFieldSummary. More...

struct	PartitionKeyEqual
	Transparent equality functor for PartitionKey with heterogeneous lookup support. More...

struct	PartitionKeyHash
	Transparent hash functor for PartitionKey with heterogeneous lookup support. More...

struct	PartitionKeyRef
	Lightweight lookup key for heterogeneous lookup without copying PartitionValues. More...

class	PartitionMap
	A map that uses a pair of spec ID and partition tuple as keys. More...

class	PartitionSet
	A set that uses a pair of spec ID and partition tuple as elements. More...

class	PartitionSpec
	A partition spec for a Table. More...

struct	PartitionStatisticsFile
	Represents a partition statistics file. More...

class	PartitionValues
	StructLike wrapper for a vector of literals that represent partition values. More...

struct	PartitionValuesEqual
	Equality functor for PartitionValues. More...

struct	PartitionValuesHash
	Hash functor for PartitionValues. More...

class	PendingUpdate
	Base class for all kinds of table metadata updates. More...

class	PositionDeleteIndex
	Tracks deleted row positions using a bitmap. More...

class	PositionDeleteWriter
	Writer for Iceberg position delete files. More...

struct	PositionDeleteWriterOptions
	Options for creating a PositionDeleteWriter. More...

class	PositionOutputStream
	Positioned byte stream for writing file contents. More...

class	PositionPathVisitor
	Visitor for building a map from field ID to position path. More...

class	Predicate
	A predicate is a boolean expression that tests a term against some criteria. More...

class	PrimitiveType
	A data type that does not have child fields. More...

class	ProjectionEvaluator
	A class that projects expressions for a table's data rows into expressions on the table's partition values, for a table's partition spec. More...

struct	Projections
	Utils to project expressions on rows to expressions on partitions. More...

class	PropertyUtil

class	PruneColumnVisitor
	Visitor for pruning columns based on selected field IDs. More...

class	Reader
	Base reader class to read data from different file formats. More...

struct	ReaderFactoryRegistry
	Registry of reader factories for different file formats. More...

struct	ReaderOptions
	Options for creating a reader. More...

class	ReaderProperties

class	Reference
	A reference represents a named field in an expression. More...

class	ReferenceVisitor
	Visitor to collect referenced field IDs from an expression. More...

class	ReplacePartitions
	API for overwriting files in a table by partition. More...

class	ResidualEvaluator
	Finds the residuals for an Expression using the partitions in the given PartitionSpec. More...

class	ResolvingFileIO
	FileIO that resolves and caches implementations by registry name. More...

struct	RetryConfig
	Configuration for retry behavior. More...

class	RetryRunner
	Utility class for running tasks with retry logic. More...

class	RewriteFiles
	API for replacing files in a table. More...

class	RewriteNot

class	RoaringPositionBitmap
	A bitmap that supports positive 64-bit positions, optimized for cases where most positions fit in 32 bits. More...

class	RollingManifestWriter
	A rolling manifest writer that can produce multiple manifest files. More...

class	RowDelta
	API for encoding row-level changes to a table. More...

class	SanitizeExpression
	Rewrites an expression tree as an unbound expression whose literal values are replaced with type-aware placeholders (e.g. "(2-digit-int)", "(hash-3f9a1c02)", "(date-5-days-ago)"). More...

class	ScanMetrics
	Live scan metrics collected during a table scan operation. More...

struct	ScanMetricsResult
	Immutable snapshot of scan metrics for use in ScanReport. More...

struct	ScanReport
	Report generated after a table scan operation. More...

class	ScanTask
	An abstract scan task. More...

class	Schema
	A schema for a Table. More...

class	SchemaCache

class	SchemaField
	A type combined with a name. More...

struct	SchemaProjection
	A schema partner to carry projection information. More...

class	ScopedLogger
	Bind a logger for the current thread until this object leaves scope. More...

class	SeekableInputStream
	Seekable byte stream for reading file contents. More...

class	SessionCatalog
	Factory interface for creating context-bound catalog views. More...

struct	SessionContext
	Session state used to bind catalog operations to a caller context. More...

class	SetSnapshot
	Sets the current snapshot directly or by rolling back. More...

struct	Snapshot
	A snapshot of the data in a table at a point in time. More...

class	SnapshotCache
	A snapshot with cached manifest loading capabilities. More...

struct	SnapshotLogEntry
	Represents a snapshot log entry. More...

class	SnapshotManager
	API for managing snapshots. More...

struct	SnapshotRef
	A reference to a snapshot, either a branch or a tag. More...

class	SnapshotsTable
	Snapshots metadata table. More...

class	SnapshotSummaryBuilder
	Helper class for building snapshot summaries. More...

struct	SnapshotSummaryFields
	Optional Snapshot Summary Fields. More...

class	SnapshotUpdate
	API for table changes that produce snapshots. More...

class	SortField
	a field with its transform. More...

class	SortOrder
	A sort order for a table. More...

struct	Split
	A split of the file to read. More...

class	StagedTable
	A table created by stage-create and not yet committed. More...

class	StaticTable
	A read-only table. More...

struct	StatisticsFile
	Represents a statistics file in the Puffin format. More...

struct	StorageCredential
	A storage credential vended for a storage location prefix. More...

class	StrictMetricsEvaluator
	Evaluates an Expression against DataFile. More...

struct	StringEqual
	Transparent equality function that supports std::string_view as lookup key. More...

struct	StringHash
	Transparent hash function that supports std::string_view as lookup key. More...

struct	StringLess
	Transparent less function that supports std::string_view as lookup key. More...

class	StringType
	A data type representing an arbitrary-length character sequence (encoded in UTF-8). More...

class	StringUtils

class	StructLike
	An immutable struct-like wrapper. More...

class	StructLikeAccessor
	An accessor for a struct-like object. More...

class	StructLikeSet
	A set of StructLike rows with type-aware hashing and equality. More...

class	StructType
	A data type representing a struct with nested fields. More...

class	SupportsStorageCredentials
	Mix-in for FileIO implementations that route object paths to per-prefix file systems built from vended storage credentials, letting the catalog stay decoupled from concrete storage implementations. More...

class	Table
	Represents an Iceberg table. More...

struct	TableIdentifier
	Identifies a table in iceberg catalog. More...

struct	TableMetadata
	Represents the metadata for an Iceberg table. More...

class	TableMetadataBuilder
	Builder class for constructing TableMetadata objects. More...

class	TableMetadataCache

struct	TableMetadataUtil
	Utility class for table metadata. More...

class	TableProperties
	Table properties for Iceberg tables. More...

class	TableRequirement
	Base class for update requirement operations. More...

class	TableRequirements
	Factory class for generating table requirements. More...

class	TableScan
	Represents a configured scan operation on a table. More...

class	TableScanBuilder
	Builder class for creating TableScan instances. More...

class	TableUpdate
	Base class for metadata update operations. More...

class	TableUpdateContext
	Context for generating table requirements. More...

struct	TaskContext
	Context passed to task creation functions. More...

class	TaskGroup

class	TemporalUtils

class	Term
	A term is an expression node that produces a typed value when evaluated. More...

class	Timer
	Abstract timer for measuring operation durations. More...

struct	TimerResult
	Serialisable snapshot of a single Timer measurement. More...

class	TimestampBase
	A base class for any timestamp time (irrespective of unit or timezone). More...

class	TimestampNsType
	A data type representing a timestamp in nanoseconds without reference to a timezone. More...

class	TimestampType
	A data type representing a timestamp in microseconds without reference to a timezone. More...

class	TimestampTzNsType
	A data type representing a timestamp as nanoseconds since the epoch in UTC. A time zone or offset is not stored. More...

class	TimestampTzType
	A data type representing a timestamp as microseconds since the epoch in UTC. A time zone or offset is not stored. More...

class	TimeType
	A data type representing a wall clock time in microseconds without reference to a timezone or date. More...

class	Transaction
	A transaction for performing multiple updates to a table. More...

class	TransactionContext
	Shared context between Transaction and PendingUpdate instances. More...

class	Transform
	Represents a transform used in partitioning or sorting in Iceberg. More...

class	TransformFunction
	A transform function used for partitioning. More...

class	TransformUtil

class	True
	An Expression that is always true. More...

class	TruncateTransform
	Truncate transform that truncates values to a specified width. More...

class	TruncateUtils

class	Type
	Interface for a data type for a field. More...

class	Unbound
	Interface for unbound expressions that need schema binding. More...

class	UnboundAggregate
	Base class for unbound aggregates. More...

class	UnboundAggregateImpl
	Template for unbound aggregates that carry a term and operation. More...

class	UnboundPredicate
	Non-template base class for all UnboundPredicate instances. More...

class	UnboundPredicateImpl
	Unbound predicates contain unbound terms and must be bound to a concrete schema before they can be evaluated. More...

class	UnboundTerm
	Base class for unbound terms. More...

class	UnboundTransform
	An unbound transform expression. More...

class	UnknownType
	A null-only placeholder type used when a more specific type is not known. More...

class	UpdateLocation
	Updating table location with a new base location. More...

class	UpdatePartitionSpec
	API for partition spec evolution. More...

class	UpdatePartitionStatistics
	Updates table partition statistics. More...

class	UpdateProperties
	Updates table properties. More...

class	UpdateSchema
	API for schema evolution. More...

class	UpdateSnapshotReference
	Updates snapshot references. More...

class	UpdateSortOrder
	Updating table sort order with a newly created order. More...

class	UpdateStatistics
	Updates table statistics. More...

class	UrlEncoder
	Utilities for encoding and decoding URLs. More...

class	Uuid

class	UuidType
	A data type representing a UUID. While defined as a distinct type, it is effectively a fixed(16). More...

class	VariantType
	A semi-structured type whose structure may vary across rows. More...

class	VoidTransform
	Void transform that discards the input and always returns null. More...

class	Writer
	Base writer class to write data from different file formats. More...

struct	WriteResult
	File metadata for files produced by a writer. More...

struct	WriterFactoryRegistry
	Registry of writer factories for different file formats. More...

struct	WriterOptions
	Options for creating a writer. More...

class	WriterProperties

class	YearTransform
	Year transform that extracts the number of years from timestamp inputs since the epoch. More...

Concepts
concept	TermType

concept	AsResult

concept	IsIncrementalScan

concept	EndianConvertible
	Concept for values that can be converted to/from another endian format.

concept	FromChars

Typedefs
using	FieldIdsSetRef = std::reference_wrapper< std::unordered_set< int32_t > >

using	ReaderFactory = std::function< Result< std::unique_ptr< Reader > >()>
	Factory function to create a reader of a specific file format.

using	WriterFactory = std::function< Result< std::unique_ptr< Writer > >()>
	Factory function to create a writer of a specific file format.

using	FatalHandler = std::function< void(const std::source_location &, std::string_view message)>
	A hook invoked on the fatal-log path just before std::abort().

using	LoggerFactory = std::function< Result< std::unique_ptr< Logger > >(const std::unordered_map< std::string, std::string > &properties)>
	Factory constructing a Logger from catalog-style properties.

using	ManifestWriterFactory = std::function< Result< std::unique_ptr< ManifestWriter > >(int32_t spec_id, ManifestContent content)>
	Factory type for creating ManifestWriter instances.

using	MetricsReport = std::variant< ScanReport, CommitReport >
	A metrics report, which can be either a ScanReport or CommitReport.

using	MetricsReporterFactory = std::function< Result< std::unique_ptr< MetricsReporter > >(const std::unordered_map< std::string, std::string > &properties)>
	Function type for creating MetricsReporter instances.

using	DurationNs = std::chrono::nanoseconds
	Duration type for metrics reporting (nanosecond precision).

using	MappedFieldConstRef = std::reference_wrapper< const MappedField >

template<typename T , typename E = typename DefaultError<T>::type>
using	Result = std::expected< T, E >
	Result alias.

using	Status = Result< void >

template<AsResult T>
using	ResultValueT = typename std::remove_cvref_t< T >::value_type

using	Scalar = std::variant< std::monostate, bool, int32_t, int64_t, float, double, std::string_view, Decimal, std::shared_ptr< StructLike >, std::shared_ptr< ArrayLike >, std::shared_ptr< MapLike > >
	A scalar value depending on its data type.

using	DataTableScanBuilder = TableScanBuilder< DataTableScan >

using	IncrementalAppendScanBuilder = TableScanBuilder< IncrementalAppendScan >

using	IncrementalChangelogScanBuilder = TableScanBuilder< IncrementalChangelogScan >

using	UncheckedStructLikeSet = StructLikeSet< false >

using	ExecutorTask = FnOnce< void()>

using	OptionalExecutor = std::optional< std::reference_wrapper< Executor > >

using	PartitionKey = std::pair< int32_t, PartitionValues >

using	TimePointMs = std::chrono::time_point< std::chrono::system_clock, std::chrono::milliseconds >
	A time point in milliseconds.

using	TimePointNs = std::chrono::time_point< std::chrono::system_clock, std::chrono::nanoseconds >
	A time point in nanoseconds.

Enumerations
enum class	LogLevel { kTrace , kDebug , kInfo , kWarn , kError , kCritical , kFatal , kOff }
	Logging severity level, ordered from most to least verbose. More...

enum class	ManifestStatus { kExisting = 0 , kAdded = 1 , kDeleted = 2 }

enum class	ManifestContent { kData = 0 , kDeletes = 1 }
	The type of files tracked by the manifest, either data or delete files; 0 for all v1 manifests. More...

enum class	CounterUnit { kCount , kBytes , kUndefined }
	Unit for a Counter metric.

enum class	MetricsReportType { kScanReport , kCommitReport }
	The type of a metrics report.

enum class	TimerUnit { kNanoseconds , kMicroseconds , kMilliseconds , kSeconds , kMinutes , kHours , kDays }
	Time units supported by Timer.

enum class	ErrorKind { kAlreadyExists , kAuthenticationFailed , kBadRequest , kCommitFailed , kCommitStateUnknown , kDecompressError , kForbidden , kInternalServerError , kInvalid , kInvalidArgument , kInvalidArrowData , kInvalidExpression , kInvalidManifest , kInvalidManifestList , kInvalidSchema , kIOError , kJsonParseError , kNamespaceNotEmpty , kNoSuchNamespace , kNoSuchPlanId , kNoSuchPlanTask , kNoSuchTable , kNoSuchWarehouse , kNoSuchView , kNotAllowed , kNotAuthorized , kNotFound , kNotImplemented , kNotSupported , kRestError , kRetryableValidationFailed , kServiceUnavailable , kTokenExpired , kUnknownError , kValidationFailed }
	Error types for iceberg.

enum class	SnapshotRefType { kBranch , kTag }
	The type of snapshot reference. More...

enum class	SortDirection { kAscending , kDescending }
	Sort direction in a partition, either ascending or descending. More...

enum class	NullOrder { kFirst , kLast }

enum class	ChangelogOperation : uint8_t { kInsert , kDelete , kUpdateBefore , kUpdateAfter }

enum class	TransactionKind : uint8_t { kCreate , kUpdate }
	Whether a transaction creates a new table or updates an existing one.

enum class	TransformType { kUnknown , kIdentity , kBucket , kTruncate , kYear , kMonth , kDay , kHour , kVoid }
	Transform types used for partitioning. More...

enum class	TypeId { kStruct , kList , kMap , kBoolean , kInt , kLong , kFloat , kDouble , kDecimal , kDate , kTime , kTimestamp , kTimestampTz , kTimestampNs , kTimestampTzNs , kString , kUuid , kFixed , kBinary , kUnknown , kVariant , kGeometry , kGeography }
	A data type. More...

enum class	TimeUnit { kMicrosecond , kNanosecond }
	The time unit. In Iceberg V3 nanoseconds are also supported.

enum class	EdgeAlgorithm { kSpherical , kVincenty , kThomas , kAndoyer , kKarney }
	The algorithm used to interpolate geography edges.

enum class	CleanupLevel : uint8_t { kNone , kMetadataOnly , kAll }
	An enum representing possible clean up levels used in snapshot expiration. More...

Functions
void ICEBERG_EXPORT	ForEachPositionDelete (std::span< const int64_t > positions, PositionDeleteIndex &target, std::vector< uint32_t > &scratch)
	Apply `positions` to `target` as deletes; semantically equivalent to calling `target.Delete(pos)` for each entry. Out-of-range positions are silently ignored. Sorted, mostly-contiguous input is fastest.

ICEBERG_EXPORT std::string_view	ToString (Expression::Operation op)
	Returns a string representation of an expression operation.

ICEBERG_EXPORT Result< Expression::Operation >	Negate (Expression::Operation op)
	Returns the negated operation.

template<typename R , typename V > requires std::derived_from<V, ExpressionVisitor<R>>
Result< R >	Visit (const std::shared_ptr< Expression > &expr, V &visitor)
	Traverse an expression tree with a visitor.

ICEBERG_EXPORT std::string_view	ToString (FileFormatType format_type)
	Convert a FileFormatType to a string.

ICEBERG_EXPORT Result< FileFormatType >	FileFormatTypeFromString (std::string_view str) noexcept
	Convert a string to a FileFormatType.

ICEBERG_EXPORT Result< bool >	GeospatialBoundsIntersect (const Type &type, const BoundingBox &lhs, const BoundingBox &rhs)
	Check whether two bounding boxes intersect for an Iceberg geospatial type.

constexpr std::string_view	ToString (LogLevel level) noexcept
	String representation of a LogLevel.

Result< LogLevel >	LogLevelFromString (std::string_view s)
	Parse a LogLevel from a string (case-insensitive).

ICEBERG_EXPORT std::shared_ptr< Logger >	GetDefaultLogger ()
	Return the process-global default logger (never null).

ICEBERG_EXPORT std::shared_ptr< Logger >	GetCurrentLogger ()
	Return the effective logger for this thread (never null): the active ScopedLogger binding if any, else the global default.

ICEBERG_EXPORT void	SetDefaultLogger (std::shared_ptr< Logger > logger)
	Install a new process-global default logger.

ICEBERG_EXPORT void	SetDefaultLevel (LogLevel level)
	Set the minimum level of the current default logger.

ICEBERG_EXPORT void	SetFatalHandler (FatalHandler handler)
	Install (or clear, with nullptr) the process-global fatal handler.

ICEBERG_EXPORT FatalHandler	GetFatalHandler ()
	Return the installed fatal handler (empty if none). Used by the fatal logging path; thread-safe.

template<typename... Args>
void	Log (LogLevel level, internal::FmtWithLoc< std::type_identity_t< Args >... > fmt, Args &&... args) noexcept
	Log to the process-default logger, std::format style. Formats only if the level is enabled; never throws.

template<typename... Args>
void	Log (Logger &logger, LogLevel level, internal::FmtWithLoc< std::type_identity_t< Args >... > fmt, Args &&... args) noexcept
	Log to an explicit logger, std::format style. Formats only if enabled.

ICEBERG_EXPORT constexpr Result< ManifestStatus >	ManifestStatusFromInt (int32_t status) noexcept
	Get the relative manifest status type from int.

ICEBERG_EXPORT constexpr std::string_view	ToString (DataFile::Content type) noexcept
	Get the relative datafile content type name.

ICEBERG_EXPORT constexpr Result< DataFile::Content >	DataFileContentFromInt (int32_t content) noexcept
	Get the relative data file content type from int.

ICEBERG_EXPORT constexpr std::string_view	ToString (ManifestContent type) noexcept
	Get the relative manifest content type name.

ICEBERG_EXPORT constexpr Result< ManifestContent >	ManifestContentFromString (std::string_view str) noexcept
	Get the relative manifest content type from name.

ICEBERG_EXPORT constexpr std::string_view	ToString (CounterUnit unit) noexcept
	String representation of a CounterUnit.

ICEBERG_EXPORT Result< CounterUnit >	CounterUnitFromString (std::string_view s)
	Parse a CounterUnit from a string.

ICEBERG_EXPORT constexpr std::string_view	ToString (MetricsReportType type) noexcept
	Get the string representation of a metrics report type.

ICEBERG_EXPORT MetricsReportType	GetReportType (const MetricsReport &report)
	Get the type of a metrics report.

ICEBERG_EXPORT constexpr std::string_view	ToString (TimerUnit unit) noexcept
	String representation of a TimerUnit.

ICEBERG_EXPORT std::string	ToString (const MappedField &field)

ICEBERG_EXPORT std::string	ToString (const MappedFields &fields)

ICEBERG_EXPORT std::string	ToString (const NameMapping &mapping)

ICEBERG_EXPORT Result< std::unique_ptr< NameMapping > >	CreateMapping (const Schema &schema)
	Create a name-based mapping for a schema.

ICEBERG_EXPORT Result< std::unique_ptr< NameMapping > >	UpdateMapping (const NameMapping &mapping, const std::unordered_map< int32_t, std::shared_ptr< SchemaField > > &updates, const std::multimap< int32_t, int32_t > &adds)
	Update a name-based mapping using changes to a schema.

Result< Scalar >	LiteralToScalar (const Literal &literal)
	Convert a Literal to a Scalar.

ICEBERG_EXPORT Result< SchemaProjection >	Project (const Schema &expected_schema, const Schema &source_schema, bool prune_source)
	Project the expected schema on top of the source schema.

ICEBERG_EXPORT std::string_view	ToString (FieldProjection::Kind kind)

ICEBERG_EXPORT std::string	ToString (const FieldProjection &projection)

ICEBERG_EXPORT std::string	ToString (const SchemaProjection &projection)

ICEBERG_EXPORT constexpr std::string_view	ToString (SnapshotRefType type) noexcept
	Get the relative snapshot reference type name.

ICEBERG_EXPORT constexpr Result< SnapshotRefType >	SnapshotRefTypeFromString (std::string_view str) noexcept
	Get the relative snapshot reference type from name.

ICEBERG_EXPORT constexpr std::string_view	ToString (SortDirection direction)
	Get the relative sort direction name.

ICEBERG_EXPORT constexpr Result< SortDirection >	SortDirectionFromString (std::string_view str)
	Get the relative sort direction from name.

ICEBERG_EXPORT constexpr std::string_view	ToString (NullOrder null_order)
	Get the relative null order name.

ICEBERG_EXPORT constexpr Result< NullOrder >	NullOrderFromString (std::string_view str)
	Get the relative null order from name.

ICEBERG_EXPORT std::string	ToString (const BlobMetadata &blob_metadata)
	Returns a string representation of a BlobMetadata.

ICEBERG_EXPORT std::string	ToString (const StatisticsFile &statistics_file)
	Returns a string representation of a StatisticsFile.

ICEBERG_EXPORT std::string	ToString (const PartitionStatisticsFile &partition_statistics_file)
	Returns a string representation of a PartitionStatisticsFile.

ICEBERG_EXPORT std::string	ToString (const Namespace &ns)

ICEBERG_EXPORT std::string	ToString (const TableIdentifier &ident)

ICEBERG_EXPORT std::string	ToString (const SnapshotLogEntry &entry)
	Returns a string representation of a SnapshotLogEntry.

ICEBERG_EXPORT std::string	ToString (const MetadataLogEntry &entry)
	Returns a string representation of a MetadataLogEntry.

ICEBERG_EXPORT constexpr std::string_view	TransformTypeToString (TransformType type)
	Get the relative transform name.

ICEBERG_EXPORT Result< std::shared_ptr< Transform > >	TransformFromString (std::string_view transform_str)
	Converts a string representation of a transform into a Transform instance.

ICEBERG_EXPORT const std::shared_ptr< BooleanType > &	boolean ()
	Return a BooleanType instance.

ICEBERG_EXPORT const std::shared_ptr< IntType > &	int32 ()
	Return an IntType instance.

ICEBERG_EXPORT const std::shared_ptr< LongType > &	int64 ()
	Return a LongType instance.

ICEBERG_EXPORT const std::shared_ptr< FloatType > &	float32 ()
	Return a FloatType instance.

ICEBERG_EXPORT const std::shared_ptr< DoubleType > &	float64 ()
	Return a DoubleType instance.

ICEBERG_EXPORT const std::shared_ptr< DateType > &	date ()
	Return a DateType instance.

ICEBERG_EXPORT const std::shared_ptr< TimeType > &	time ()
	Return a TimeType instance.

ICEBERG_EXPORT const std::shared_ptr< TimestampType > &	timestamp ()
	Return a TimestampType instance.

ICEBERG_EXPORT const std::shared_ptr< TimestampTzType > &	timestamp_tz ()
	Return a TimestampTzType instance.

ICEBERG_EXPORT const std::shared_ptr< TimestampNsType > &	timestamp_ns ()
	Return a TimestampNsType instance.

ICEBERG_EXPORT const std::shared_ptr< TimestampTzNsType > &	timestamptz_ns ()
	Return a TimestampTzNsType instance.

ICEBERG_EXPORT const std::shared_ptr< BinaryType > &	binary ()
	Return a BinaryType instance.

ICEBERG_EXPORT const std::shared_ptr< StringType > &	string ()
	Return a StringType instance.

ICEBERG_EXPORT const std::shared_ptr< UuidType > &	uuid ()
	Return a UuidType instance.

ICEBERG_EXPORT const std::shared_ptr< UnknownType > &	unknown ()
	Return an UnknownType instance.

ICEBERG_EXPORT const std::shared_ptr< VariantType > &	variant ()
	Return a VariantType instance.

ICEBERG_EXPORT const std::shared_ptr< GeometryType > &	geometry ()
	Return the default GeometryType instance.

ICEBERG_EXPORT const std::shared_ptr< GeographyType > &	geography ()
	Return the default GeographyType instance.

ICEBERG_EXPORT std::shared_ptr< DecimalType >	decimal (int32_t precision, int32_t scale)
	Create a DecimalType with the given precision and scale.

ICEBERG_EXPORT std::shared_ptr< FixedType >	fixed (int32_t length)
	Create a FixedType with the given length.

ICEBERG_EXPORT std::shared_ptr< GeometryType >	geometry (std::string crs)
	Create a GeometryType with the given CRS.

ICEBERG_EXPORT std::shared_ptr< GeographyType >	geography (std::string crs)
	Create a GeographyType with the given CRS.

ICEBERG_EXPORT std::shared_ptr< GeographyType >	geography (std::string crs, EdgeAlgorithm algorithm)
	Create a GeographyType with the given CRS and edge algorithm.

ICEBERG_EXPORT std::shared_ptr< StructType >	struct_ (std::vector< SchemaField > fields)
	Create a StructType with the given fields.

ICEBERG_EXPORT std::shared_ptr< ListType >	list (SchemaField element)
	Create a ListType with the given element field.

ICEBERG_EXPORT std::shared_ptr< MapType >	map (SchemaField key, SchemaField value)
	Create a MapType with the given key and value fields.

ICEBERG_EXPORT std::string_view	ToString (TypeId id)
	Get the lowercase string representation of a TypeId.

ICEBERG_EXPORT std::string_view	ToString (EdgeAlgorithm algorithm)
	Get the lowercase string representation of an EdgeAlgorithm.

ICEBERG_EXPORT Result< EdgeAlgorithm >	EdgeAlgorithmFromString (std::string_view name)
	Parse a lowercase edge algorithm name.

ICEBERG_EXPORT std::ostream &	operator<< (std::ostream &os, const Decimal &decimal)

template<EndianConvertible T>
constexpr T	ByteSwap (T value)
	Byte-swap a value. For floating-point types, only support 32-bit and 64-bit floats.

template<EndianConvertible T>
constexpr T	ToLittleEndian (T value)
	Convert a value to little-endian format.

template<EndianConvertible T>
constexpr T	FromLittleEndian (T value)
	Convert a value from little-endian format.

template<EndianConvertible T>
constexpr T	ToBigEndian (T value)
	Convert a value to big-endian format.

template<EndianConvertible T>
constexpr T	FromBigEndian (T value)
	Convert a value from big-endian format.

template<EndianConvertible T>
void	WriteLittleEndian (T value, void *output)
	Write a value in little-endian format to a buffer.

template<EndianConvertible T>
T	ReadLittleEndian (const void *input)
	Read a value in little-endian format from a buffer.

template<EndianConvertible T>
void	WriteBigEndian (T value, void *output)
	Write a value in big-endian format to a buffer.

template<EndianConvertible T>
T	ReadBigEndian (const void *input)
	Read a value in big-endian format from a buffer.

ICEBERG_EXPORT auto	MakeCommitRetryRunner (int32_t num_retries, int32_t min_wait_ms, int32_t max_wait_ms, int32_t total_timeout_ms)
	Helper function to create a RetryRunner with table commit configuration.

ICEBERG_EXPORT Result< bool >	StructLikeEqual (const StructLike &lhs, const StructLike &rhs)
	Compare two StructLike rows by scalar values.

ICEBERG_EXPORT TimePointMs	TimePointMsFromUnixMs (int64_t unix_ms)
	Returns a TimePointMs from a Unix timestamp in milliseconds.

ICEBERG_EXPORT int64_t	UnixMsFromTimePointMs (TimePointMs time_point_ms)
	Returns a Unix timestamp in milliseconds from a TimePointMs.

ICEBERG_EXPORT TimePointNs	TimePointNsFromUnixNs (int64_t unix_ns)
	Returns a TimePointNs from a Unix timestamp in nanoseconds.

ICEBERG_EXPORT int64_t	UnixNsFromTimePointNs (TimePointNs time_point_ns)
	Returns a Unix timestamp in nanoseconds from a TimePointNs.

ICEBERG_EXPORT std::string	FormatTimePointMs (TimePointMs time_point_ms)
	Returns a human-readable string representation of a TimePointMs.

ICEBERG_EXPORT TimePointMs	CurrentTimePointMs ()
	Returns a time point in milliseconds that represents the current system time.

ICEBERG_EXPORT std::unordered_map< int32_t, int32_t >	IndexParents (const StructType &root_struct)
	Index parent field IDs for all fields in a struct hierarchy.

ICEBERG_EXPORT Result< std::shared_ptr< Schema > >	AssignFreshIds (int32_t schema_id, const Schema &schema, std::function< int32_t()> next_id)
	Assigns fresh IDs to all fields in a schema.

ICEBERG_EXPORT bool	IsPromotionAllowed (const std::shared_ptr< Type > &from_type, const std::shared_ptr< Type > &to_type)
	Check if type promotion from one type to another is allowed.

template<typename VISITOR , typename... ARGS>
Status	VisitTypeInline (const Type &type, VISITOR *visitor, ARGS &&... args)
	Calls `visitor` with the corresponding concrete type class.

template<typename VISITOR , typename... ARGS>
auto	VisitType (const Type &type, VISITOR &&visitor, ARGS &&... args) -> decltype(std::forward< VISITOR >(visitor)(type, args...))
	Call `visitor` with the corresponding concrete type class.

template<typename VISITOR , typename... ARGS>
Status	VisitTypeIdInline (TypeId id, VISITOR *visitor, ARGS &&... args)
	Calls `visitor` with a nullptr of the corresponding concrete type class.

template<typename VISITOR , typename... ARGS>
auto	VisitTypeCategory (const Type &type, VISITOR *visitor, ARGS &&... args)
	Visit a type using a categorical visitor pattern.

Variables
constexpr std::string_view	kParquetFieldIdKey = "PARQUET:field_id"

constexpr int64_t	kInvalidSnapshotId = -1

constexpr int64_t	kInvalidSequenceNumber = -1

constexpr int64_t	kInvalidSchemaId = -1

constexpr int64_t	kUnassignedSequenceNumber = -1
	Stand-in for the current sequence number that will be assigned when the commit is successful. This is replaced when writing a manifest list by the ManifestFile adapter.

constexpr std::string_view	kLevelProperty = "level"
	Well-known Logger::Initialize() property keys.

constexpr std::string_view	kPatternProperty = "pattern"

constexpr std::string_view	kLoggerImpl = "logger-impl"
	Property key selecting the logger implementation.

constexpr std::string_view	kLoggerTypeNoop = "noop"
	Built-in logger type identifiers.

constexpr std::string_view	kLoggerTypeCerr = "cerr"

constexpr std::string_view	kLoggerTypeSpdlog = "spdlog"

constexpr std::string_view	kMetricsReporterImpl = "metrics-reporter-impl"
	Property key for configuring the metrics reporter implementation.

constexpr std::string_view	kMetricsReporterTypeNoop = "noop"
	Property value for the noop metrics reporter.

template class ICEBERG_EXTERN_TEMPLATE_CLASS_EXPORT	TableScanBuilder< DataTableScan >

template class ICEBERG_EXTERN_TEMPLATE_CLASS_EXPORT	TableScanBuilder< IncrementalAppendScan >

template class ICEBERG_EXTERN_TEMPLATE_CLASS_EXPORT	TableScanBuilder< IncrementalChangelogScan >

constexpr size_t	kHashPrime = 0x9e3779b9

template class ICEBERG_EXTERN_TEMPLATE_CLASS_EXPORT	StructLikeSet< true >

template class ICEBERG_EXTERN_TEMPLATE_CLASS_EXPORT	StructLikeSet< false >

Detailed Description

Core Apache Iceberg C++ APIs.

Typedef Documentation

◆ FatalHandler

using iceberg::FatalHandler = typedef std::function<void(const std::source_location&, std::string_view message)>

A hook invoked on the fatal-log path just before std::abort().

Receives the fatal record's source location and already-formatted message. Runs after the record has been emitted and the logger flushed, so an embedder (JNI, a Python host, a crash reporter) can print a stack trace, flush its own resources, or translate the abort. It must not return normally on the expectation of cancelling the abort – ICEBERG_LOG_FATAL always terminates; if the handler itself does not exit the process, std::abort() still runs after it.

◆ MetricsReport

using iceberg::MetricsReport = typedef std::variant<ScanReport, CommitReport>

A metrics report, which can be either a ScanReport or CommitReport.

This variant type allows handling both report types uniformly through the MetricsReporter interface.

◆ MetricsReporterFactory

using iceberg::MetricsReporterFactory = typedef std::function<Result<std::unique_ptr<MetricsReporter> >( const std::unordered_map<std::string, std::string>& properties)>

Function type for creating MetricsReporter instances.

Parameters

properties Configuration properties for the reporter.

Returns: A new MetricsReporter instance or an error.

◆ Scalar

using iceberg::Scalar = typedef std::variant<std::monostate, bool, int32_t, int64_t, float, double, std::string_view, Decimal, std::shared_ptr<StructLike>, std::shared_ptr<ArrayLike>, std::shared_ptr<MapLike> >

A scalar value depending on its data type.

Note that all string and binary values are stored as non-owned string_view and their lifetime is managed by the wrapped object.

Enumeration Type Documentation

◆ CleanupLevel

enum class iceberg::CleanupLevel : uint8_t

strong

An enum representing possible clean up levels used in snapshot expiration.

Enumerator
kNone	Skip all file cleanup, only remove snapshot metadata.
kMetadataOnly	Clean up only metadata files (manifests, manifest lists, statistics), retain data files.
kAll	Clean up both metadata and data files (default).

◆ LogLevel

enum class iceberg::LogLevel

strong

Logging severity level, ordered from most to least verbose.

Levels are ordered so that level >= threshold is the enabled test. kOff is the maximum sentinel: as a threshold it disables all emission (it is never the level of an actual message).

◆ ManifestContent

enum class iceberg::ManifestContent

strong

The type of files tracked by the manifest, either data or delete files; 0 for all v1 manifests.

Enumerator
kData	The manifest content is data.
kDeletes	The manifest content is deletes.

◆ NullOrder

enum class iceberg::NullOrder

strong

Enumerator
kFirst	Nulls are sorted first.
kLast	Nulls are sorted last.

◆ SnapshotRefType

enum class iceberg::SnapshotRefType

strong

The type of snapshot reference.

Enumerator
kBranch	Branches are mutable named references that can be updated by committing a new snapshot as the branch’s referenced snapshot using the Commit Conflict Resolution and Retry procedures.
kTag	Tags are labels for individual snapshots.

◆ SortDirection

enum class iceberg::SortDirection

strong

Sort direction in a partition, either ascending or descending.

Enumerator
kAscending	Ascending.
kDescending	Descending.

◆ TransformType

enum class iceberg::TransformType

strong

Transform types used for partitioning.

Enumerator
kUnknown	Used to represent some customized transform that can't be recognized or supported now.
kIdentity	Equal to source value, unmodified.
kBucket	Hash of value, mod `N`
kTruncate	Value truncated to width `W`
kYear	Extract a date or timestamp year, as years from 1970.
kMonth	Extract a date or timestamp month, as months from 1970-01.
kDay	Extract a date or timestamp day, as days from 1970-01-01.
kHour	Extract a timestamp hour, as hours from 1970-01-01 00:00:00.
kVoid	Always produces `null`

◆ TypeId

enum class iceberg::TypeId

strong

A data type.

This is not a complete data type by itself because some types are nested and/or parameterized.

Iceberg V3's unknown type is supported as a null-only placeholder type.

Function Documentation

◆ AssignFreshIds()

ICEBERG_EXPORT Result< std::shared_ptr< Schema > > iceberg::AssignFreshIds	(	int32_t	schema_id,
		const Schema &	schema,
		std::function< int32_t()>	next_id
	)

Assigns fresh IDs to all fields in a schema.

Parameters

schema_id	An ID assigned to this schema
schema	The schema to assign IDs to.
next_id	An id assignment function, which returns the next ID to assign.

Returns: A schema with new ids assigned by the next_id function.

◆ CounterUnitFromString()

ICEBERG_EXPORT Result< CounterUnit > iceberg::CounterUnitFromString ( std::string_view s )

inline

Parse a CounterUnit from a string.

Parameters

s	The string to parse ("count", "bytes", or "undefined").

Returns: The CounterUnit, or an InvalidArgument error if unrecognized.

◆ CreateMapping()

ICEBERG_EXPORT Result< std::unique_ptr< NameMapping > > iceberg::CreateMapping ( const Schema & schema )

Create a name-based mapping for a schema.

The mapping returned by this method will use the schema's name for each field.

Parameters

schema The schema to create the mapping for.

Returns: A new NameMapping instance initialized with the schema's fields and names.

◆ ForEachPositionDelete()

void ICEBERG_EXPORT iceberg::ForEachPositionDelete	(	std::span< const int64_t >	positions,
		PositionDeleteIndex &	target,
		std::vector< uint32_t > &	scratch
	)

Apply positions to target as deletes; semantically equivalent to calling target.Delete(pos) for each entry. Out-of-range positions are silently ignored. Sorted, mostly-contiguous input is fastest.

Parameters

scratch Caller-owned reusable buffer for the bulk dispatch path. Cleared and reused per key group; retain across calls to amortize allocations. Pass a distinct scratch per thread when calling concurrently with disjoint target.

◆ GeospatialBoundsIntersect()

ICEBERG_EXPORT Result< bool > iceberg::GeospatialBoundsIntersect	(	const Type &	type,
		const BoundingBox &	lhs,
		const BoundingBox &	rhs
	)

Check whether two bounding boxes intersect for an Iceberg geospatial type.

Returns an error for invalid boxes or a non-geospatial type.

◆ GetCurrentLogger()

ICEBERG_EXPORT std::shared_ptr< Logger > iceberg::GetCurrentLogger ( )

Return the effective logger for this thread (never null): the active ScopedLogger binding if any, else the global default.

Off the hot path – returns an owning copy, e.g. to capture the current logger and re-bind it on a worker thread (see ScopedLogger). During teardown, prefer the Log(...) overloads over emitting through this handle.

◆ GetDefaultLogger()

ICEBERG_EXPORT std::shared_ptr< Logger > iceberg::GetDefaultLogger ( )

Return the process-global default logger (never null).

Off the hot path – acquires the slot lock and returns an owning copy. The logging path uses the cheaper internal hot-path accessor instead.

◆ GetReportType()

ICEBERG_EXPORT MetricsReportType iceberg::GetReportType ( const MetricsReport & report )

inline

Get the type of a metrics report.

Parameters

report The metrics report to get the type of.

Returns: The type of the metrics report.

◆ IndexParents()

ICEBERG_EXPORT std::unordered_map< int32_t, int32_t > iceberg::IndexParents ( const StructType & root_struct )

Index parent field IDs for all fields in a struct hierarchy.

Parameters

root_struct The root struct type to analyze

Returns: A map from field ID to its parent struct field ID

Note

This function assumes the input StructType has already been validated:

All field IDs must be non-negative
All field IDs must be unique across the entire schema hierarchy If the struct is part of a Schema, these invariants are enforced by StructType::InitFieldById which checks for duplicate field IDs.

◆ IsPromotionAllowed()

ICEBERG_EXPORT bool iceberg::IsPromotionAllowed	(	const std::shared_ptr< Type > &	from_type,
		const std::shared_ptr< Type > &	to_type
	)

Check if type promotion from one type to another is allowed.

Type promotion rules:

unknown -> any primitive type
int -> long
float -> double
decimal(P,S) -> decimal(P',S) where P' > P

Parameters

from_type	The original type
to_type	The target type

Returns: true if promotion is allowed, false otherwise

◆ Log() [1/2]

template<typename... Args>

void iceberg::Log	(	Logger &	logger,
		LogLevel	level,
		internal::FmtWithLoc< std::type_identity_t< Args >... >	fmt,
		Args &&...	args
	)

noexcept

Log to an explicit logger, std::format style. Formats only if enabled.

Example: iceberg::Log(logger, LogLevel::kWarn, "retry {}", attempt);

◆ Log() [2/2]

template<typename... Args>

void iceberg::Log	(	LogLevel	level,
		internal::FmtWithLoc< std::type_identity_t< Args >... >	fmt,
		Args &&...	args
	)

noexcept

Log to the process-default logger, std::format style. Formats only if the level is enabled; never throws.

Example: iceberg::Log(LogLevel::kInfo, "loaded {} files", n);

◆ LogLevelFromString()

Result< LogLevel > iceberg::LogLevelFromString ( std::string_view s )

inline

Parse a LogLevel from a string (case-insensitive).

Parameters

s	The string to parse ("trace", "debug", "info", "warn", "error", "critical", "fatal", or "off").

Returns: The LogLevel, or an InvalidArgument error if unrecognized.

◆ Project()

ICEBERG_EXPORT Result< SchemaProjection > iceberg::Project	(	const Schema &	expected_schema,
		const Schema &	source_schema,
		bool	prune_source
	)

Project the expected schema on top of the source schema.

Parameters

expected_schema	The expected schema.
source_schema	The source schema.
prune_source	Whether the source schema can be pruned to project the expected schema on it. For example, literally a Parquet reader implementation is capable of column pruning, so `prune_source` is set to true in this case such that the `from` field in `FieldProjection` exactly reflects the position (relative to its nesting level) to get the column value from the reader.

Returns: The projection result.

◆ ReadBigEndian()

template<EndianConvertible T>

T iceberg::ReadBigEndian ( const void * input )

Read a value in big-endian format from a buffer.

Note: Caller must ensure input has at least sizeof(T) bytes available.

◆ ReadLittleEndian()

template<EndianConvertible T>

T iceberg::ReadLittleEndian ( const void * input )

Read a value in little-endian format from a buffer.

Note: Caller must ensure input has at least sizeof(T) bytes available.

◆ SetDefaultLevel()

ICEBERG_EXPORT void iceberg::SetDefaultLevel ( LogLevel level )

Set the minimum level of the current default logger.

Convenience for GetDefaultLogger()->SetLevel(level). Filtering is always decided by the logger's own ShouldLog(), so changing a logger's level by any means (this, SetLevel on a held handle, or Initialize) takes effect immediately.

◆ SetDefaultLogger()

ICEBERG_EXPORT void iceberg::SetDefaultLogger ( std::shared_ptr< Logger > logger )

Install a new process-global default logger.

A null argument installs the no-op logger. Thread-safe; intended for occasional (configuration-time) use rather than the hot path.

◆ SetFatalHandler()

ICEBERG_EXPORT void iceberg::SetFatalHandler ( FatalHandler handler )

Install (or clear, with nullptr) the process-global fatal handler.

Thread-safe. Intended to be set once at startup. Replaces any previous handler.

◆ ToString()

ICEBERG_EXPORT std::string_view iceberg::ToString ( TypeId id )

Get the lowercase string representation of a TypeId.

This returns the same lowercase string as used by Type::ToString() methods. For example: TypeId::kBoolean -> "boolean", TypeId::kInt -> "int", etc.

Parameters

id	The TypeId to convert to string

Returns: A string_view containing the lowercase type name

◆ TransformFromString()

ICEBERG_EXPORT Result< std::shared_ptr< Transform > > iceberg::TransformFromString ( std::string_view transform_str )

Converts a string representation of a transform into a Transform instance.

This function parses the provided string to identify the corresponding transform type (e.g., "identity", "year", "bucket[16]"), and creates a shared pointer to the corresponding Transform object. It supports both simple transforms (like "identity") and parameterized transforms (like "bucket[16]" or "truncate[4]").

Parameters

transform_str The string representation of the transform type.

Returns: A Result containing either a shared pointer to the corresponding Transform instance or an Error if the string does not match any valid transform type.

◆ UpdateMapping()

ICEBERG_EXPORT Result< std::unique_ptr< NameMapping > > iceberg::UpdateMapping	(	const NameMapping &	mapping,
		const std::unordered_map< int32_t, std::shared_ptr< SchemaField > > &	updates,
		const std::multimap< int32_t, int32_t > &	adds
	)

Update a name-based mapping using changes to a schema.

Parameters

mapping	a name-based mapping
updates	a map from field ID to updated field definitions
adds	a map from parent field ID to nested fields to be added

Returns: an updated mapping with names added to renamed fields and the mapping extended for new fields

◆ Visit()

template<typename R , typename V >
requires std::derived_from<V, ExpressionVisitor<R>>

Result< R > iceberg::Visit	(	const std::shared_ptr< Expression > &	expr,
		V &	visitor
	)

Traverse an expression tree with a visitor.

This function traverses the given expression tree in postorder traversal and calls appropriate visitor methods for each node. Results from child nodes are passed to parent nodes.

Template Parameters

R	The return type produced by the visitor
V	The visitor type (must derive from ExpressionVisitor<R>)

Parameters

expr	The expression to traverse
visitor	The visitor to use for traversal

Returns: The result produced by the visitor for the root expression node

◆ VisitType()

template<typename VISITOR , typename... ARGS>

auto iceberg::VisitType	(	const Type &	type,
		VISITOR &&	visitor,
		ARGS &&...	args
	)		-> decltype(std::forward<VISITOR>(visitor)(type, args...))

inline

Call visitor with the corresponding concrete type class.

Template Parameters

ARGS	Additional arguments, if any, will be passed to the Visit function after the `type` argument

Unlike VisitTypeInline which calls visitor.Visit, here visitor itself is called. visitor must support a const Type& argument as a fallback, in addition to concrete type classes.

The intent is for this to be called on a generic lambda that may internally use if constexpr or similar constructs.

◆ VisitTypeCategory()

template<typename VISITOR , typename... ARGS>

auto iceberg::VisitTypeCategory	(	const Type &	type,
		VISITOR *	visitor,
		ARGS &&...	args
	)

inline

Visit a type using a categorical visitor pattern.

This function provides a simplified visitor interface that groups Iceberg types into five categories based on their structural properties:

Struct types: Complex types with named fields (StructType)
List types: Sequential container types (ListType)
Map types: Key-value container types (MapType)
Variant type: Semi-structured type that is neither nested nor primitive (VariantType)
Primitive types: All leaf types without nested structure (primitive types)

This grouping is useful for algorithms that need to distinguish between container types and leaf types, but don't require separate handling for each primitive type variant (e.g., Int vs Long vs String).

Template Parameters

VISITOR	Visitor class that must implement five Visit methods: `VisitStruct(const StructType&, ARGS...)` for struct types `VisitList(const ListType&, ARGS...)` for list types `VisitMap(const MapType&, ARGS...)` for map types `VisitVariant(const VariantType&, ARGS...)` for the variant type `VisitPrimitive(const PrimitiveType&, ARGS...)` for all primitive types
ARGS	Additional argument types forwarded to Visit methods

Parameters

type	The type to visit
visitor	Pointer to the visitor instance
args	Additional arguments forwarded to the Visit methods

Returns: The return value from the invoked Visit method

◆ VisitTypeIdInline()

template<typename VISITOR , typename... ARGS>

Status iceberg::VisitTypeIdInline	(	TypeId	id,
		VISITOR *	visitor,
		ARGS &&...	args
	)

inline

Calls visitor with a nullptr of the corresponding concrete type class.

Template Parameters

VISITOR	Visitor type that implements Visit() for all Iceberg types.
ARGS	Additional arguments, if any, will be passed to the Visit function after the `type` argument

Returns: Status

◆ VisitTypeInline()

template<typename VISITOR , typename... ARGS>

Status iceberg::VisitTypeInline	(	const Type &	type,
		VISITOR *	visitor,
		ARGS &&...	args
	)

inline

Calls visitor with the corresponding concrete type class.

Template Parameters

VISITOR	Visitor type that implements Visit() for all Iceberg types.
ARGS	Additional arguments, if any, will be passed to the Visit function after the `type` argument

Returns: Status

A visitor is a type that implements specialized logic for each Iceberg type. Example usage:

class ExampleVisitor {
  Status Visit(const IntType& type) { ... }
  Status Visit(const LongType& type) { ... }
  ...
}
ExampleVisitor visitor;
VisitTypeInline(some_type, &visitor);

◆ WriteBigEndian()

template<EndianConvertible T>

void iceberg::WriteBigEndian	(	T	value,
		void *	output
	)

Write a value in big-endian format to a buffer.

Note: Caller must ensure output has at least sizeof(T) bytes available.

◆ WriteLittleEndian()

template<EndianConvertible T>

void iceberg::WriteLittleEndian	(	T	value,
		void *	output
	)

Write a value in little-endian format to a buffer.

Note: Caller must ensure output has at least sizeof(T) bytes available.

Variable Documentation

◆ kLevelProperty

constexpr std::string_view iceberg::kLevelProperty = "level"

inlineconstexpr

Well-known Logger::Initialize() property keys.

level is honored by the base Logger::Initialize (parsed via LogLevelFromString) on every backend. pattern is honored only by the spdlog backend; CerrLogger uses a fixed layout and ignores it.

◆ kMetricsReporterImpl

constexpr std::string_view iceberg::kMetricsReporterImpl = "metrics-reporter-impl"

constexpr

Property key for configuring the metrics reporter implementation.

Set this property in catalog properties to specify which metrics reporter implementation to use. The value should match a registered reporter type.

Namespaces

Classes

Concepts

Typedefs

Enumerations

Functions

Variables

Detailed Description

Typedef Documentation

◆ FatalHandler

◆ MetricsReport

◆ MetricsReporterFactory

◆ Scalar

Enumeration Type Documentation

◆ CleanupLevel

◆ LogLevel

◆ ManifestContent

◆ NullOrder

◆ SnapshotRefType

◆ SortDirection

◆ TransformType

◆ TypeId

Function Documentation

◆ AssignFreshIds()

◆ CounterUnitFromString()

◆ CreateMapping()

◆ ForEachPositionDelete()

◆ GeospatialBoundsIntersect()

◆ GetCurrentLogger()

◆ GetDefaultLogger()

◆ GetReportType()

◆ IndexParents()

◆ IsPromotionAllowed()

◆ Log() [1/2]

◆ Log() [2/2]

◆ LogLevelFromString()

◆ Project()

◆ ReadBigEndian()

◆ ReadLittleEndian()

◆ SetDefaultLevel()

◆ SetDefaultLogger()

◆ SetFatalHandler()

◆ ToString()

◆ TransformFromString()

◆ UpdateMapping()

◆ Visit()

◆ VisitType()

◆ VisitTypeCategory()

◆ VisitTypeIdInline()

◆ VisitTypeInline()

◆ WriteBigEndian()

◆ WriteLittleEndian()

Variable Documentation

◆ kLevelProperty

◆ kMetricsReporterImpl