nontabular

Nontabular is an adjective used to describe data, representations, or formats that are not arranged in a fixed two-dimensional table. It contrasts with tabular data, which is organized into rows and columns. Nontabular data encompasses unstructured data such as text, audio, and video; semi-structured data like JSON or XML; and various organized forms such as graphs, hierarchical trees, and multidimensional arrays.

In computing and data management, tabular data is often associated with relational databases, while nontabular data is typically stored in non-relational systems. Document stores hold JSON-like documents, graph databases model entities as nodes and relationships as edges, and column-family stores manage sparse columns. Data lakes and object stores preserve raw nontabular data in its native form. Querying nontabular data commonly requires different models and tools, including document queries, graph traversals, or schema-on-read approaches.

Applications and processing methods for nontabular data frequently involve transforming it into a tabular representation for analysis, feature extraction, or reporting. This can enable use with traditional machine learning algorithms. Conversely, some modern models, such as natural language processing or graph neural networks, can operate directly on certain nontabular formats without full flattening.

The term appears in discussions of data modeling, integration, and governance to emphasize data that does not fit a simple table schema. It is part of a broader taxonomy that also includes tabular and semi-structured data.