datarecord

A data record is a coherent collection of related data fields that together describe a single instance of an entity within a dataset. It is the basic unit of information in many data storage and processing systems, such as databases, data warehouses, and streaming pipelines. Each record represents an object, event, or observation and is intended to be independently meaningful.

A data record consists of fields or attributes, each with a name, type, and value. It often carries a unique identifier or primary key to distinguish it from other records. Metadata such as source, creation timestamp, and lineage may accompany the record to aid provenance and governance.

In relational databases, a record corresponds to a row in a table. In NoSQL and document stores, a record may be a document with nested fields. On disk or in transit, records are serialized in formats like JSON, XML, CSV, Avro, or Parquet, depending on performance and interoperability requirements.

Records are usually defined by a schema that specifies allowed fields and types. Schema evolution allows fields to be added or retired over time. Records are processed individually or in streams, enabling row-level analytics, filtering, enrichment, and joining with other records.

Maintaining data quality for records involves validation, completeness checks, consistency constraints, and privacy controls. Data management practices emphasize accurate identification, lineage tracking, and secure handling to ensure reliable downstream analytics.