DFDL (Data Format Description Language) is a modeling language, standardized by the Open Grid Forum, for describing general text and binary data formats in a standard way so they can be parsed into and serialized from a common data model without changing the original files.
DFDL
DFDL (Data Format Description Language) is a standard, published by the Open Grid Forum, for describing the layout of arbitrary text and binary data formats. Rather than inventing a new file format, DFDL uses a subset of XML Schema annotated with format details so that existing data — fixed-width records, delimited files, industry message formats — can be read into and written back out from a common logical model. The original bytes are never changed; DFDL simply teaches a parser how to interpret them.
- XML Schema based - Reuses XML Schema element and type structures, adding annotations that describe the physical layout.
- Bidirectional - The same description drives both parsing (bytes to data) and unparsing (data back to bytes).
- Format neutral - Handles text, binary, delimited, and fixed-length data, including legacy and industry-specific formats.
- Non-invasive - Describes data in place without requiring conversion to a new storage format.
In API and integration work, DFDL matters wherever legacy or non-JSON data has to meet modern interfaces. It shows up in middleware and integration platforms that need to lift EDI, financial messaging, or fixed-width mainframe records into a structured model an API can serve, and Apache Daffodil provides an open-source reference implementation. As more of that translation gets automated, a machine-readable description of a format — like DFDL provides — is what lets tooling parse messy real-world data reliably instead of hand-coding a parser for each one.