What cardinality analysis is
Measure how many values are distinct
Cardinality compares the distinct values in a field with all present values observed at the same normalized array path. It helps distinguish compact categories from highly unique data.
Data modeling
Separate dimensions from highly varied fields
Low-cardinality fields often contain statuses, regions, or flags. High-cardinality fields contain mostly distinct values. UtilityDock reports the measurement without prescribing a model.
Identifier review
Locate fields with one value per record
Complete string or number fields with 100% cardinality are marked as candidate identifiers. Confirm their business meaning separately before treating them as keys.
ETL and analytics
Profile dimensions before importing data
Cardinality can reveal whether a field is suitable for grouping, whether repeated values dominate, and whether missing data affects the observed population.
Examples
Read each percentage in context
| Result | Meaning |
|---|---|
| 2 of 100 · 2% | Low-cardinality field with two repeated categories. |
| 60 of 100 · 60% | Moderate-cardinality field with repeated and distinct values. |
| 100 of 100 · 100% | High-cardinality field and possible identifier candidate. |
Inspection utilities
Continue with related UtilityDock inspection workflows
FAQ
Questions about JSON field cardinality
Does UtilityDock upload or store my JSON?
No. Your files never leave your computer. All processing happens locally inside your browser.
What is field cardinality?
Field cardinality is the number of distinct present values. The percentage divides that count by all present values for the same field.
How are low- and high-cardinality fields classified?
Low cardinality is 10% or less distinct among present values. High cardinality is 90% or more. Fields between those thresholds are moderate.
What is a candidate identifier field?
A candidate is a complete string or number field with one distinct value per record. This is a measured signal, not a guarantee of business uniqueness.
Are missing fields counted as values?
No. Missing properties are reported separately. Explicit null remains a present JSON value and participates in distinct counting.