AI Glossary / Data engineering and architecture / Advanced
Data item
Definition
A data item is an identifiable unit of information that a system treats as a value or structured part of its data. The exact meaning depends on context: it may refer to a field, an occurrence of a value or a named component within a record. Some programming languages also distinguish elementary items from group items. The term should therefore be accompanied by enough context to show what is being counted or described.
What it means
A quantity of five can be a value stored in an order's Quantity field. The field's definition explains what the value means, while the particular occurrence belongs to a specific order. Calling both a data item without clarification can blur that distinction. A larger structure, such as an address containing street and city components, may also be treated as an item by a program. The important question is which unit the current operation handles and how that unit relates to its surrounding information.
How it works
A system identifies an item through its structure, name, position or another defined reference and associates it with a representation and meaning. In IBM's COBOL documentation, an elementary item has no subordinate data items, while a group item contains subordinate items. That is one precise language-specific model, not a universal definition for every database or file format. When documenting a general workflow, specify whether the item is a single field value, a structured group or a complete record before describing operations on it.
Three practical examples
- A form stores a supplied delivery quantity as one value associated with a named field and an order identifier.
- A program handles an address as a group while still allowing access to its city component.
- A data-quality report identifies which individual value failed a rule instead of vaguely saying that the entire dataset is wrong.
Think of it this way
Imagine a labeled box on a paper form. Its label describes the kind of information expected, while the writing inside is one person's response. A section containing several boxes may also be handled as a unit when the form is copied or reviewed. Data items have a similar context dependence. To communicate clearly, specify whether you mean the definition of the box, its particular contents or the larger section that contains it.
Worked example
Consider an original table with two order records. Each record contains Order ID, Quantity and Unit Price, giving three field definitions and six field-value occurrences. The quantities are 2 and 5; these are two occurrences of the same defined field. If a process updates the second quantity from 5 to 6, it changes one occurrence, not the definition of Quantity and not all six values. Calling this a one-item update is meaningful only after stating that item means one field-value occurrence here. A different system might treat the complete record as its update unit, so the wording must match the actual operation.
Illustrative case study
A fictional integration team receives an instruction to correct the customer data item. One developer interprets it as a name field, while another expects a complete customer object. The team replaces the vague instruction with a precise record identifier, component name and intended new value, then checks the result at that scope. The scenario illustrates why a broad term needs a clear referent. It does not imply that every update should be performed at field level or that a system's physical write unit must match the logical unit described by a user.
Common misconceptions and mistakes
A data item is not necessarily one byte or one character. Its representation may occupy many storage units or contain subordinate components. The name of an item is also different from its current value. Another mistake is assuming that the same value in two records represents the same entity or occurrence. Context and identifiers matter. Finally, a count of items is ambiguous unless the counting unit is defined: fields, populated cells, records and nested objects can produce different legitimate counts for the same underlying collection.
Business use
The concept is useful when specifying data requirements, corrections and validation results. Replace vague item references with an explicit unit and location in the logical structure. Record field meanings and types separately from particular values, and explain whether a proposed change affects one occurrence or the shared definition. For AI-generated structured outputs, make the expected unit clear so a response containing a whole object is not confused with one extracted field. Precise language helps reviewers verify the requested operation and recognize unintended changes in scope.
People also ask
Is a data item always the smallest possible stored unit?
No. It may be structured or contain subordinate items.
Is a field definition the same as one field value?
No. The definition describes the field; a value is one occurrence under that definition.
Why can two reports give different item counts?
They may count different units, such as records and individual field values.
Check your understanding
How many field definitions did the toy table have?
Answer: Three.
How many field-value occurrences did it contain?
Answer: Six.
What changed when the second quantity moved from 5 to 6?
Answer: One occurrence.
Why did the example define its use of item explicitly?
Answer: Other contexts may use the term for a different unit of data.
Related terms
Learn next
Take a small record and identify its field definitions, current values and any grouped components. Describe a proposed correction in three ways: as a field-value change, a record replacement and a definition change. Explain how their consequences differ, even when casual conversation might call each a change to a data item.
More in Data engineering and architecture
- Data auditing - Data auditing is a systematic examination of data and, where relevant, the processes that handle it against stated criteria.
- Data binding - Data binding is a mechanism that connects data values to another representation so that values can be displayed, transferred or updated according to defined...
- Data blending - Data blending combines information from separate sources for an analysis or view, using agreed links and compatible levels of detail.
- Data-centric computing - Data-centric computing is an approach to computer-system design that treats the location, movement and access patterns of data as central concerns.
- Data-centric programming language - A data-centric programming language emphasizes describing, querying and transforming collections of data.
- Data-centric security - Data-centric security is an approach that organizes protection around the information itself: what it contains, how sensitive it is, who may use it and...
Sources
IBM, Using data items and group items: https://www.ibm.com/docs/en/cobol-zos/6.3?topic=constants-using-data-items-group-items The order-table calculation, form analogy and integration-team case are original educational illustrations. Indexed IBM documentation supplies the elementary/group distinction in COBOL; the entry explicitly avoids treating that language-specific terminology as the only meaning of data item in every information system.
Previous: Data-intensive computing Next: Data lake