RDF file

1. Purpose

The RDF file dataset plugin reads RDF data from a local file (or ZIP archive) into the project and, for supported formats, can also write RDF data back to a file.

Typical use cases:

• Loading small to medium RDF files (e.g. .ttl, .nt, .rdf, .nq) into a project.
• Providing a local RDF snapshot for exploration, mapping, or linking.
• Exporting RDF data from a workflow into a simple N-Triples file.

Important: The entire dataset is loaded in memory. Very large files should be imported into an external RDF store and accessed via a SPARQL dataset instead.

2. Input and output

Input

• Source:
  • A single RDF file, or
  • A ZIP archive containing multiple RDF files.

The file is read from the configured resource; no upstream operators are required.

Output

• Data: RDF graph(s) exposed as a dataset that can be:
  • queried (e.g. via SPARQL-based components),
  • used as input for transformations,
  • or used in linking tasks.
• Graphs:
  • If a graph name is configured and supported by the chosen format, only that named graph is exposed.
  • Otherwise, the default graph is used.

3. Configuration notes

The detailed parameter list is generated in the UI. This section explains how to think about the most relevant settings.

3.1 File

• Points to the RDF file or ZIP archive to be used as the dataset.
• If the file is a ZIP archive:
  • all contained files are scanned,
  • only files matching the ZIP file regex (see below) are loaded.

3.2 Format

• Optional parameter that specifies the RDF serialization format (e.g. RDF/XML, N-Triples, N-Quads, Turtle).
• If left empty:
  • the format is auto-detected from the file name/extension,
  • if no format can be detected, an error is raised and the dataset will not load.
• Reading:
  • Supported formats include at least RDF/XML, N-Triples, N-Quads, and Turtle.
• Writing:
  • Only N-Triples is supported for writing.
  • For other formats, the dataset is effectively read-only.

3.3 Graph

• Optional graph name (IRI) that tells the dataset which graph to expose.
• Behaviour:
  • For graph-aware formats (e.g. N-Quads):
    • you must provide a graph name to select which graph is used.
  • For graph-less formats (e.g. N-Triples, Turtle):
    • the graph parameter is ignored and the data is treated as a single default graph.
• If left empty:
  • the default graph of the loaded dataset is used.

3.4 Entity list (advanced)

• Advanced parameter that allows restricting which entities (subjects) are retrieved.
• You can provide a list of entity IRIs, separated by whitespace or line breaks.
• Typical use:
  • limit retrieval to a known set of resources during exploration,
  • avoid scanning the entire file when only specific entities are needed.

3.5 ZIP file regex (advanced)

• Only relevant if the configured file is a ZIP archive.
• Regular expression that filters which files inside the ZIP are considered.
  • Default: .* (all files).
• Examples:
  • .*\.ttl → only Turtle files.
  • data-.*\.nt → only N-Triples files whose names start with data-.

4. Behaviour

When the dataset is used (e.g. queried or read by downstream components), it behaves as follows:

1. File size check

   • Before loading, the file size is checked against an internal maximum for in-memory processing.
   • If the file is too large, the dataset fails with a “resource too large” error rather than attempting to load it.

2. Loading RDF data

   • The RDF file (or all matching files inside a ZIP) is parsed into an internal RDF dataset held in memory.
   • The dataset may contain:
     • a default graph,
     • and, depending on the format, one or more named graphs.

3. Graph selection

   • If a graph name is configured and the selected format supports named graphs, that graph is used.
   • Otherwise, the default graph is used.
   • For graph-less formats (e.g. N-Triples, Turtle), the graph setting is ignored.

4. Data access

   • Once loaded, the data can be:
     • queried for types and paths,
     • used to retrieve entities by schema or by URI,
     • sampled for values and schema extraction.
   • Repeated reads re-use the in-memory dataset as long as the underlying file has not changed.

5. File changes

   • The dataset monitors changes to the underlying file:
     • if the file’s modification timestamp changes,
     • the RDF data is reloaded the next time it is accessed.

6. Writing

   • When used as a sink (writing RDF):
     • data is serialized as N-Triples to the configured file.
     • other output formats are not supported.

5. Limitations and recommendations

• In-memory dataset

  • All data is held in memory; this is suitable for:
    • small to medium RDF files,
    • test datasets,
    • local snapshots.
  • For large production datasets, prefer:
    • importing the data into an external RDF store,
    • and using a SPARQL dataset instead.

• Format restrictions

  • Reading: multiple standard RDF formats are supported.
  • Writing: only N-Triples is available as output format.

• Graph handling

  • Graph selection only works where the underlying format supports named graphs.
  • For formats without named graphs, the graph parameter is ignored.

6. Examples

6.1 Simple Turtle file as dataset

• File: data/example.ttl
• Format: (empty → auto-detects Turtle)
• Graph: (empty → default graph)

Use case:

• Small RDF dataset loaded from a Turtle file for exploration and transformations.

6.2 N-Quads file with explicit graph

• File: data/dump.nq
• Format: N-Quads
• Graph: http://example.com/graph/main

Use case:

• Large export containing multiple graphs, but only a specific named graph is used in the project.

6.3 ZIP archive with multiple RDF files

• File: snapshots/rdf-data.zip
• Format: (empty → formats are detected per file)
• ZIP file regex: .*\.ttl

Use case:

• ZIP archive containing various files; only Turtle files are loaded into the dataset.

7. When to use this dataset

Use the RDF file dataset when:

• You have a local RDF file or ZIP archive you want to work with directly.
• The data volume fits comfortably into memory.
• You need a simple, file-based way to:
  • explore RDF data,
  • run mappings or link specifications,
  • or export RDF as N-Triples.

For large or frequently updated datasets, prefer an external RDF store and a SPARQL dataset.

Parameter

File

The RDF file. This may also be a zip archive of multiple RDF files.

• ID: file
• Datatype: resource
• Default Value: None

Format

Optional RDF format. If left empty, it will be auto-detected based on the file extension. N-Triples is the only format that can be written, while other formats can only be read.

• ID: format
• Datatype: string
• Default Value: None

Graph

The graph name to be read. If not provided, the default graph will be used. Must be provided if the format is N-Quads.

• ID: graph
• Datatype: string
• Default Value: None

Advanced Parameter

Entity list

A list of entities to be retrieved. If not given, all entities will be retrieved. Multiple entities are separated by whitespace.

• ID: entityList
• Datatype: multiline string
• Default Value: None

ZIP file regex

If the input resource is a ZIP file, files inside the file are filtered via this regex.

• ID: zipFileRegex
• Datatype: string
• Default Value: .*

Comments