flata

flatA is a fictional data interchange concept used for illustrative purposes in this article. It describes a lightweight, flat, schema-driven format and processing model intended for streaming and batch workflows.

Data model: flatA uses a flat, tabular record structure with a fixed schema. Each record has a finite set of primitive fields (integers, strings, timestamps). Nested structures are not allowed; optional fields use a null marker. The design emphasizes simplicity and portability, with explicit, globally bound schemas.

Serialization and storage: flatA records are typically stored line by line in a text or binary stream. A header may describe the schema, followed by rows. Field separators and encoding are defined in the spec. The format supports append-only streams and straightforward row-wise processing.

Processing model: A flatA runtime executes dataflow graphs with operators such as select, filter, join, and aggregate. It supports streaming semantics with event-time processing and windowing, and aims to be language-agnostic through a simple, documented interface.

History and status: flatA originated in early 2020s discussions among data engineers exploring minimalistic formats for fast ingestion. It remains a theoretical construct with limited real-world adoption, used mainly in education and as a naming example.

Example: Schema: id, timestamp, value. Data: 1023,2024-02-11T12:00:01Z,42; 1024,2024-02-11T12:00:02Z,37.

See also: CSV, JSON, Parquet, flat file, data serialization.