Export Schema

Each Export Operation writes a table to the linked enrichment datastore. The columns depend on the asset type. This page lists every column included in each export.

Quality Checks

The Quality Check export captures one row per check on the selected containers.

Column	Type	Notes
`container`	object	Container name and datastore stub
`template`	object \| null	Parent template reference, when the check was created from a template
`fields`	list	Fields the check targets, with type and tags
`description`	string \| null
`rule_type`	string	The rule the check enforces (e.g. `isNotNull`, `isUnique`)
`coverage`	number	Expected coverage of the check
`inferred`	boolean	`true` when the check was created by AI Managed Checks
`filter`	string \| null	SQL filter applied before evaluation
`properties`	object	Rule-type-specific configuration
`deleted_at`	timestamp \| null	Non-null only for archived checks
`global_tags`	list	Tags attached to the check
`last_editor`	object \| null	The user who last modified the check
`last_updated`	timestamp
`first_asserted`	timestamp \| null	The earliest scan that produced a `pass` or `fail` result. Preserved across out-of-order partition scans (it never moves forward).
`last_asserted`	timestamp \| null	The most recent scan that produced a `pass` or `fail` result.
`is_passing`	boolean \| null	The check's current pass/fail status
`weight`	number	Weight used in quality scoring
`additional_metadata`	object	Free-form metadata attached to the check
`status`	string	`Active`, `Draft`, `Invalid`, or `Discarded`

Both first_asserted and last_asserted reflect the scan start time at the moment the check produced a result, in UTC. They are null for checks that have never been asserted.

Anomalies

The Anomaly export captures one row per anomaly on the selected containers.

Column	Type	Notes
`container`	object	Container name and datastore stub
`uuid`	string	Stable identifier, preserved across re-runs and scan iterations
`type`	string	`shape` or `record`
`status`	string	`Active`, `Acknowledged`, `Resolved`, `Invalid`, `Duplicate`, or `Discarded`
`global_tags`	list	Tags attached to the anomaly

The Anomaly export is intentionally minimal so it can be joined with the platform's anomaly API for additional detail when needed.

Field Profiles

The Field Profile export captures one row per (container, field) pair from the most recent profile.

Column	Type	Notes
`field_id`	integer	ID of the profiled field
`field_name`	string	Field name
`field_display_name`	string \| null	Display name of the field, when one is set
`field_native_type`	string \| null	The field's type as declared in the source system (e.g. `VARCHAR`)
`field_status`	string	`active`, `masked`, `missing`, or `excluded`. See the note below the table.
`field_description`	string \| null	Field description
`field_quality_score`	number \| null	Field-level quality score. Currently always exported as `null`; read field quality scores in the platform or through the API instead.
`field_weight`	number \| null	Field weight used in quality scoring
`field_global_tags`	string	Comma-separated tag names on the field
`datastore_id`	integer	Source datastore ID
`source_datastore`	string	Source datastore name
`datastore_type`	string	Source datastore connector type
`container_id`	integer	Source container ID
`source_container`	string	Source container name
`container_container_type`	string	`table`, `file`, `computed_table`, etc.
`name`	string	Field name
`field_type`	string	`STRING`, `INT`, `FLOAT`, `BOOLEAN`, `TIMESTAMP`, etc.
`type_declared`	boolean	Whether the type was explicitly declared in the source schema
`completeness`	number \| null	Fraction of non-null values, 0..1
`approximate_distinct_values`	number \| null	HyperLogLog estimate of distinct values
`mean`	number \| null
`min`	number \| null
`max`	number \| null
`sum`	number \| null
`std_dev`	number \| null	Standard deviation
`kurtosis`	number \| null
`skewness`	number \| null
`entropy`	number \| null	Shannon entropy
`min_length`	number \| null
`max_length`	number \| null
`unique_distinct_ratio`	number \| null
`is_not_normal`	boolean \| null	Normality test outcome
`q1`	number \| null	First quartile
`median`	number \| null	Second quartile
`q3`	number \| null	Third quartile
`kll`	object \| null	KLL quantile sketch (used to derive quantile queries)
`histogram_buckets`	list \| null	List of `{value, count}` pairs. The `value` is masked when the field is under field masking and `include_masked` was not set. See Field Masking.

Because the export keeps one row per field from the most recent profile, it can include fields that are no longer present in the source as the schema evolves. Use field_status to tell them apart: active and masked fields are still operational, missing fields were not found in the latest profile results, and excluded fields were manually removed from analysis. See Field Status Types for details on each status.