#include <schema.h>

Inheritance diagram for iceberg::Schema:

Public Member Functions
	Schema (std::vector< SchemaField > fields, int32_t schema_id=kInitialSchemaId)

int32_t	schema_id () const
	Get the schema ID.

std::string	ToString () const override
	Get a user-readable string representation.

Result< std::optional< std::reference_wrapper< const SchemaField > > >	FindFieldByName (std::string_view name, bool case_sensitive=true) const
	Recursively find the SchemaField by field name.

Result< std::optional< std::reference_wrapper< const SchemaField > > >	FindFieldById (int32_t field_id) const
	Recursively find the SchemaField by field id.

Result< std::optional< std::string_view > >	FindColumnNameById (int32_t field_id) const
	Returns the canonical field name for the given id.

Result< std::unique_ptr< StructLikeAccessor > >	GetAccessorById (int32_t field_id) const
	Get the accessor to access the field by field id.

Result< std::unique_ptr< Schema > >	Select (std::span< const std::string > names, bool case_sensitive=true) const
	Creates a projected schema from selected field names.

Result< std::unique_ptr< Schema > >	Project (const std::unordered_set< int32_t > &field_ids) const
	Creates a projected schema from selected field IDs.

const std::vector< int32_t > &	IdentifierFieldIds () const
	Return the field IDs of the identifier fields.

Result< std::vector< std::string > >	IdentifierFieldNames () const
	Return the canonical field names of the identifier fields.

Result< int32_t >	HighestFieldId () const
	Get the highest field ID in the schema.

bool	SameSchema (const Schema &other) const
	Checks whether this schema is equivalent to another schema while ignoring the schema id.

Status	Validate (int32_t format_version) const
	Validate the schema for a given format version.

Public Member Functions inherited from iceberg::StructType
	StructType (std::vector< SchemaField > fields)

TypeId	type_id () const override
	Get the type ID.

std::string	ToString () const override
	Get a user-readable string representation.

std::span< const SchemaField >	fields () const override
	Get a view of the child fields.

Result< std::optional< SchemaFieldConstRef > >	GetFieldById (int32_t field_id) const override
	Get a field by field ID.

Result< std::optional< SchemaFieldConstRef > >	GetFieldByIndex (int32_t index) const override
	Get a field by index.

Result< std::optional< SchemaFieldConstRef > >	GetFieldByName (std::string_view name, bool case_sensitive) const override
	Get a field by name. Return an error Status if the field name is not unique; prefer GetFieldById or GetFieldByIndex when possible.

std::unique_ptr< Schema >	ToSchema () const

virtual Result< std::optional< SchemaFieldConstRef > >	GetFieldByName (std::string_view name, bool case_sensitive) const=0
	Get a field by name. Return an error Status if the field name is not unique; prefer GetFieldById or GetFieldByIndex when possible.

Result< std::optional< SchemaFieldConstRef > >	GetFieldByName (std::string_view name) const
	Get a field by name (case-sensitive).

Public Member Functions inherited from iceberg::NestedType
bool	is_primitive () const override
	Is this a primitive type (may not have child fields)?

bool	is_nested () const override
	Is this a nested type (may have child fields)?

Result< std::optional< SchemaFieldConstRef > >	GetFieldByName (std::string_view name) const
	Get a field by name (case-sensitive).

Static Public Member Functions
static Result< std::unique_ptr< Schema > >	Make (std::vector< SchemaField > fields, int32_t schema_id, std::vector< int32_t > identifier_field_ids)
	Create a schema.

static Result< std::unique_ptr< Schema > >	Make (std::vector< SchemaField > fields, int32_t schema_id, const std::vector< std::string > &identifier_field_names)
	Create a schema.

static Status	ValidateIdentifierFields (int32_t field_id, const Schema &schema, const std::unordered_map< int32_t, int32_t > &id_to_parent)
	Validate that the identifier field with the given ID is valid for the schema.

static const std::shared_ptr< Schema > &	EmptySchema ()
	Get an empty schema.

Static Public Attributes
static constexpr int32_t	kInitialSchemaId = 0

static constexpr int32_t	kInitialColumnId = 0

static constexpr int32_t	kInvalidColumnId = -1

static constexpr std::string_view	kAllColumns = "*"
	Special value to select all columns from manifest files.

Static Public Attributes inherited from iceberg::StructType
static constexpr TypeId	kTypeId = TypeId::kStruct

Friends
bool	operator== (const Schema &lhs, const Schema &rhs)

Additional Inherited Members
Public Types inherited from iceberg::NestedType
using	SchemaFieldConstRef = std::reference_wrapper< const SchemaField >

Protected Member Functions inherited from iceberg::StructType
bool	Equals (const Type &other) const override
	Compare two types for equality.

Static Protected Member Functions inherited from iceberg::StructType
static Result< std::unordered_map< int32_t, SchemaFieldConstRef > >	InitFieldById (const StructType &)

static Result< std::unordered_map< std::string_view, SchemaFieldConstRef > >	InitFieldByName (const StructType &)

static Result< std::unordered_map< std::string, SchemaFieldConstRef > >	InitFieldByLowerCaseName (const StructType &)

Protected Attributes inherited from iceberg::StructType
std::vector< SchemaField >	fields_

Lazy< InitFieldById >	field_by_id_

Lazy< InitFieldByName >	field_by_name_

Lazy< InitFieldByLowerCaseName >	field_by_lowercase_name_

Detailed Description

A schema for a Table.

A schema is a list of typed columns, along with a unique integer ID. A Table may have different schemas over its lifetime due to schema evolution.

Member Function Documentation

◆ EmptySchema()

const std::shared_ptr< Schema > & iceberg::Schema::EmptySchema ( )

static

Get an empty schema.

An empty schema has no fields and a schema ID of 0.

◆ FindColumnNameById()

Result< std::optional< std::string_view > > iceberg::Schema::FindColumnNameById ( int32_t field_id ) const

Returns the canonical field name for the given id.

Parameters

field_id The id of the field to get the canonical name for.

Returns: The canocinal column name of the field with the given id, or std::nullopt if not found.

◆ FindFieldById()

Result< std::optional< std::reference_wrapper< const SchemaField > > > iceberg::Schema::FindFieldById ( int32_t field_id ) const

Recursively find the SchemaField by field id.

Parameters

field_id The id of the field to get the accessor for.

Returns: The field with the given id, or std::nullopt if not found.

◆ FindFieldByName()

Result< std::optional< std::reference_wrapper< const SchemaField > > > iceberg::Schema::FindFieldByName	(	std::string_view	name,
		bool	case_sensitive = `true`
	)		const

Recursively find the SchemaField by field name.

Short names for maps and lists are included for any name that does not conflict with a canonical name. For example, a list, 'l', of structs with field 'x' will produce short name 'l.x' in addition to canonical name 'l.element.x'. A map 'm', if its value includes a struct with field 'x' will produce short name 'm.x' in addition to canonical name 'm.value.x'. FIXME: Currently only handles ASCII lowercase conversion; extend to support non-ASCII characters (e.g., using std::towlower or ICU)

◆ GetAccessorById()

Result< std::unique_ptr< StructLikeAccessor > > iceberg::Schema::GetAccessorById ( int32_t field_id ) const

Get the accessor to access the field by field id.

Parameters

field_id The id of the field to get the accessor for.

Returns: The accessor to access the field, or NotFound if the field is not found.

◆ HighestFieldId()

Result< int32_t > iceberg::Schema::HighestFieldId ( ) const

Get the highest field ID in the schema.

Returns: The highest field ID.

◆ Make() [1/2]

Result< std::unique_ptr< Schema > > iceberg::Schema::Make	(	std::vector< SchemaField >	fields,
		int32_t	schema_id,
		const std::vector< std::string > &	identifier_field_names
	)

static

Create a schema.

Parameters

fields	The fields that make up the schema.
schema_id	The unique identifier for this schema (default: kInitialSchemaId).
identifier_field_names	Canonical names of fields that uniquely identify rows in the table.

Returns: A new Schema instance or Status if failed.

◆ Make() [2/2]

Result< std::unique_ptr< Schema > > iceberg::Schema::Make	(	std::vector< SchemaField >	fields,
		int32_t	schema_id,
		std::vector< int32_t >	identifier_field_ids
	)

static

Create a schema.

Parameters

fields	The fields that make up the schema.
schema_id	The unique identifier for this schema (default:kInitialSchemaId).
identifier_field_ids	Field IDs that uniquely identify rows in the table.

Returns: A new Schema instance or Status if failed.

◆ Project()

Result< std::unique_ptr< Schema > > iceberg::Schema::Project ( const std::unordered_set< int32_t > & field_ids ) const

Creates a projected schema from selected field IDs.

Parameters

field_ids Set of field IDs to select

Returns: Projected schema containing only the specified fields.

Note: Field ID of a nested field may not be projected unless at least one of its sub-fields has been projected.

◆ schema_id()

int32_t iceberg::Schema::schema_id ( ) const

Get the schema ID.

A schema is identified by a unique ID for the purposes of schema evolution.

◆ Select()

Result< std::unique_ptr< Schema > > iceberg::Schema::Select	(	std::span< const std::string >	names,
		bool	case_sensitive = `true`
	)		const

Creates a projected schema from selected field names.

Parameters

names	Selected field names and nested names are dot-concatenated.
case_sensitive	Whether name matching is case-sensitive (default: true).

Returns: Projected schema containing only selected fields.

Note: If the field name of a nested type has been selected, all of its sub-fields will be selected.

◆ ToString()

std::string iceberg::Schema::ToString ( ) const

overridevirtual

Get a user-readable string representation.

Implements iceberg::util::Formattable.

◆ Validate()

Status iceberg::Schema::Validate ( int32_t format_version ) const

Validate the schema for a given format version.

This validates that the schema does not contain types that were released in later format versions.

Parameters

format_version The format version to validate against.

Returns: Error status if the schema is invalid.

◆ ValidateIdentifierFields()

Status iceberg::Schema::ValidateIdentifierFields	(	int32_t	field_id,
		const Schema &	schema,
		const std::unordered_map< int32_t, int32_t > &	id_to_parent
	)

static

Validate that the identifier field with the given ID is valid for the schema.

This method checks that the specified field ID represents a valid identifier field according to Iceberg's identifier field requirements. It verifies that the field:

exists in the schema
is a primitive type
is not optional (required field)
is not a float or double type
is not nested within optional or non-struct parent fields

Parameters

field_id	The ID of the field to validate as an identifier field.
schema	The schema containing the field to validate.
id_to_parent	A mapping from field IDs to their parent field IDs for nested field validation.

Returns: Status indicating success or failure of the validation.

The documentation for this class was generated from the following files:

iceberg/schema.h
iceberg/schema.cc

Public Member Functions

Static Public Member Functions

Static Public Attributes

Friends

Additional Inherited Members

Detailed Description

Member Function Documentation

◆ EmptySchema()

◆ FindColumnNameById()

◆ FindFieldById()

◆ FindFieldByName()

◆ GetAccessorById()

◆ HighestFieldId()

◆ Make() [1/2]

◆ Make() [2/2]

◆ Project()

◆ schema_id()

◆ Select()

◆ ToString()

◆ Validate()

◆ ValidateIdentifierFields()