TypeDB 3.0 is live! Get started for free.

Output formats

At present, TypeDB supports two different output formats for data pipelines.

If you would like this list to be extended with your favorite data formats, drop us a message on our Discord community server!

Concept rows

Concept rows are TypeDB’s native data format. All data pipeline stages (except for the fetch stage) take streams of concepts as inputs and output streams of concept rows themselves.

A concept row is an association of named variables (the “column names”) to concepts, i.e., type labels, data instances, or literal values. A concept row may be written of the form

Schematic format of concept rows
{ $var_1=concept, $var_2=concept, ..., $var_n }

When working with optional variables in data pipelines, concept assignment may be left empty, indicated by (), e.g.:

A concept row with an empty (third) variable
{ $var_1=type_label, $var_2=3.1415, $var_3=(), $var_4=0x001 }

JSON

JSON outputs can be produced using a fetch stage in a data pipeline. They represent documents (sometimes referenced as concept documents in TypeDB) with custom literal keys and values containing other documents, lists, or TypeDB concepts: types, instances, and values.