eu.fbk.rdfpro

Interface Mapper

Functional Interface:

This is a functional interface and can therefore be used as the assignment target for a lambda expression or method reference.

@FunctionalInterface
public interface Mapper

Mapper function in a MapReduce job.

A Mapper object is used in a MapReduce job (see RDFProcessors.mapReduce(Mapper, Reducer, boolean)) for mapping a statement to zero or more Value keys; statements are then grouped by key and each partition is processed in a reduce job (see Reducer) to produce the final output statements. Mapping is done by method map(Statement). It can return zero keys to drop the statement, or a key array including BYPASS_KEY to force the statement to be directly emitted in output, bypassing the reduce stage.

A common type of Mapper is the one extracting a component or hashing a subset of components of the input statement. This mapper is already implemented and can be instantiated using the factory method select(String).

Implementations of this interface should be thread-safe, as method map() is meant to be invoked concurrently by different threads on different statements.

Field Summary

Fields

Modifier and Type	Field and Description
static org.openrdf.model.URI	BYPASS_KEY Special key used to bypass the reduce stage and directly emit the statement in output.

Method Summary

Modifier and Type	Method and Description
static Mapper	bypass(Mapper mapper, Predicate<org.openrdf.model.Statement> predicate) Returns a bypassed version of the input mapper that skips and marks with BYPASS_KEY quads matching the specified predicate.
static Mapper	concat(Mapper... mappers) Returns a Mapper returning a concatenation of all the keys produced by the Mappers supplied for the input statement.
static Mapper	filter(Mapper mapper, Predicate<org.openrdf.model.Statement> predicate) Returns a filtered version of the input mapper that only maps quads matching the supplied predicate.
org.openrdf.model.Value[]	map(org.openrdf.model.Statement statement) Maps a statement to zero or more Value keys.
static Mapper	parse(String expression) Parses a Mapper out of the supplied expression string.
static Mapper	select(String components) Returns a Mapper returning a single key based on one or more selected components of the input statement.

Field Detail

BYPASS_KEY

static final org.openrdf.model.URI BYPASS_KEY

Special key used to bypass the reduce stage and directly emit the statement in output.

Method Detail

map

org.openrdf.model.Value[] map(org.openrdf.model.Statement statement)
                       throws org.openrdf.rio.RDFHandlerException

Maps a statement to zero or more Value keys. When used in a MapReduce job, returning zero keys has the effect of dropping the statements, otherwise the statement will be put in each key partition, each one being later processed in a reduce job (see Reducer); if one of the keys is BYPASS_KEY, then the statement will be (also) directly emitted in output, bypassing the reduce stage (this can be used to apply the MapReduce paradigm to only a subset of statements). Null keys are admissible in output (e.g., to denote the default context).

Parameters:

statement - the statement to map, not null

Returns:

an array with zero or more Value keys associated to the statement, not null

Throws:

org.openrdf.rio.RDFHandlerException - on error

bypass

static Mapper bypass(Mapper mapper,
                     @Nullable
                     Predicate<org.openrdf.model.Statement> predicate)

Returns a bypassed version of the input mapper that skips and marks with BYPASS_KEY quads matching the specified predicate.

Parameters:

mapper - the mapper to filter

predicate - the predicate; if null, no bypassing is performed

Returns:

the resulting mapper

filter

static Mapper filter(Mapper mapper,
                     @Nullable
                     Predicate<org.openrdf.model.Statement> predicate)

Returns a filtered version of the input mapper that only maps quads matching the supplied predicate.

Parameters:

mapper - the mapper to filter

predicate - the predicate; if null, no filtering is performed

Returns:

the resulting mapper

concat

static Mapper concat(Mapper... mappers)

Returns a Mapper returning a concatenation of all the keys produced by the Mappers supplied for the input statement. Duplicate keys for the same statement are merged.

Parameters:

mappers - the mappers whose output has to be concatenated

Returns:

the created Mapper

select

static Mapper select(String components)

Returns a Mapper returning a single key based on one or more selected components of the input statement. Parameter components is a s , p, o, c string specifying which components should be selected. If a single component is selected it is returned as the key unchanged. If multiple components are selected, they are merged and hashed to produce the returned key. In any case, exactly one key is returned for each input statement.

Parameters:

components - a string of symbols s , p, o, c specifying which components to select, not null nor empty

Returns:

the created Mapper

parse

@Nullable
static Mapper parse(@Nullable
                              String expression)

Parses a Mapper out of the supplied expression string. The expression can be a language: expression script or a component expression supported by select(String).

Parameters:

expression - the expression to parse

Returns:

the parsed mapper, or null if a null expression was supplied