Rowrecord

Rowrecord is a term used to describe a data unit that represents a single row in a tabular dataset or database table. In practice, a rowrecord encapsulates the values for all columns in that row, typically as a mapping from column identifiers to values. Depending on the context, it may be referred to as a row, a record, or a row-wise data item.

Structure and representation often vary by language and framework. A rowrecord can be implemented as a dictionary or map, a simple object or struct with fields for each column, or an immutable record with a fixed schema. In streaming and ETL pipelines, each rowrecord passes through processing stages, carrying the row’s field values intact.

Usage and contexts include relational databases, CSV or TSV parsers, data processing frameworks, APIs that emit tabular data, and spreadsheet applications. When a schema is defined for the dataset, a rowrecord adheres to that schema, with consistent field names and data types. Null or missing values are commonly allowed depending on the system’s rules for missing data.

Schema and typing considerations vary: some environments enforce a fixed schema for rowrecords, enabling stronger type checks, while others use dynamic typing. Serialization formats for rowrecords include JSON, XML, Avro, and Protobuf, and in storage systems, row-oriented formats preserve full row data for row-wise access, as opposed to columnar storage that optimizes columnar access patterns.

See also row-based storage, record, dataset, table.