Skip to main content

Master Data Management

Master Data Management lets a space maintain governed reference and master datasets inside Qarion. Use it for data that many products, policies, contracts, reports, or teams rely on: country codes, currencies, cost centers, legal entities, product hierarchies, customer segments, consent categories, and similar lookup tables.

Master data is catalog-backed. It has owners, tags, lifecycle metadata, comments, attachments, versions, export destinations, and audit events instead of living only in a spreadsheet.

Feature gate

Master Data pages require the master_data.enabled feature.

When To Use​

Use Master Data when a dataset should be edited, reviewed, published, and reused as a controlled business asset.

  • Use a reference dataset for lookup values such as ISO countries, risk ratings, regions, or code lists.
  • Use a master dataset for controlled business entities such as customers, suppliers, products, departments, or cost centers.
  • Use ordinary data products when the asset is primarily discovered, documented, or governed from an upstream system rather than maintained directly in Qarion.

Space And Organization Views​

Master Data is scoped to spaces. The space view at /master-data and /spaces/{spaceSlug}/master-data shows datasets in the active or selected space and allows dataset creation when the tier limit has not been reached.

The organization view aggregates datasets across spaces in the selected organization. It is a discovery view: it shows the owning space on each dataset card and routes you into that space's dataset detail page. Creation remains space-scoped.

Dataset List​

The Master Data landing page supports search, category filtering, pagination, and category cards for reference and master datasets. Each card shows:

  • dataset name and summary
  • category
  • up to three tags
  • current published version
  • row count
  • owning space in organization view
  • governed styling when the dataset uses approval-required publishing

Create A Dataset​

From the Master Data list, select New Dataset and provide:

  • Name: required; Qarion generates a unique space-scoped slug from it.
  • Description: optional short summary.
  • Category: reference or master.
  • Governance mode: direct_edit or approval_required.
  • Tags: optional comma-separated tags.

The creator becomes the dataset owner. A new dataset starts at version 0 until the first publish creates version 1.

Dataset Detail Tabs​

Open a dataset to work across these tabs:

TabUse it for
OverviewDescription, detailed Markdown notes, catalog metadata, tags, and schema columns.
DataSpreadsheet-style row editing, CSV/XLSX import, and publish or submit actions.
ContractsData contracts where the master dataset is a producer or consumer.
VersionsPublished version history and diffs between snapshots.
ExportsExport destinations and delivery history.
GovernanceGovernance mode and role assignments.
ReferencesPublished document pages that explicitly mention the dataset.
DiscussionComments, replies, search, and file attachments embedded into the discussion.

Schema Editing​

Define columns from the Overview tab before adding rows, then update the schema as the dataset evolves. Supported column types are:

  • string
  • integer
  • decimal
  • boolean
  • date
  • datetime
  • enum

Each column can include a description, field sensitivity, PII/PHI flag, and constraints for required, unique, and primary-key fields. Enum columns can reference a source dataset slug so the published reference list can be used as the allowed value set.

Schema edits are live metadata changes. Publish a new version after schema or row changes when consumers need an immutable snapshot and version history.

Data Editing And Import​

The Data tab loads the live rows for the dataset and lets editors work in a spreadsheet-style grid:

  • Click a cell to edit it inline.
  • Use the bottom placeholder row to add a new row.
  • Delete rows from the row action.
  • Save happens as rows are added or edited.
  • The publish state remains dirty until the changes are published or submitted.

Use Import to upload .csv or .xlsx files. Imports use the first row as headers, match headers to dataset column names, apply case-insensitive matches when possible, coerce values to the target column type, warn about unmapped file columns or missing required columns, and insert valid rows. Current import limits are 25 MB per upload and 50,000 parsed rows.

Publishing And Versions​

Publishing creates an immutable point-in-time snapshot of the current schema and rows.

Direct-change datasets show Publish. Approval-required datasets show Submit and route through the configured governance workflow when one exists. Both flows require a change summary and can include an optional version label.

Published versions:

  • increment current_version
  • record row count, change summary, optional label, publisher, and publish time
  • preserve the schema and rows as they existed at publish time
  • can be compared from the Versions tab
  • can be marked with breaking-change details when schema changes are detected

Foreign-key and enum validations run during publish. Foreign-key behavior depends on the dataset's FK violation mode: warning mode appends warnings to the change summary, while blocking mode prevents publish until violations are resolved.

Drafts And Approval Workflows​

Qarion supports named dataset drafts and merge workflows at the API/service layer for parallel editing and conflict handling. The shipped dataset detail page currently centers users on the live Data tab plus publish or submit actions. If your space uses approval-required governance, use the Governance tab to switch the dataset to Approval Required and assign the responsible owner, steward, custodian, DPO, or data architect roles before users submit changes.

Governance​

The Governance tab controls how changes move to a published version:

  • Direct Change lets authorized editors publish directly.
  • Approval Required requires changes to be submitted for review before publishing.

Assign governance roles to users or teams. Supported role types are owner, steward, custodian, DPO, and data architect. Governance assignments help route accountability for access, stewardship, privacy, technical custody, and model design standards.

Master Data also participates in catalog metadata workflows. The Overview tab includes maintenance status, criticality, sensitivity, lifecycle stage, environment, and tags so the dataset can be scored by the documentation dashboard and evaluated by policies.

Contracts, References, Comments, And Attachments​

Master datasets can participate in data contracts as producers or consumers. Use the Contracts tab to review linked contracts.

The References tab shows published document pages that mention the dataset. Use this when a dataset has runbooks, definitions, or governance notes in Documents.

The Discussion tab supports threaded comments, search, and file uploads embedded into Markdown comments. Master data attachments are stored against the dataset and follow the shared Qarion attachment controls for file type, file size, download, and deletion.

Exports​

Use the Exports tab to configure destinations for published dataset versions. Supported export types are:

  • PostgreSQL with connection string, table name, and primary key column.
  • S3 / GCS with bucket, key prefix, and json or csv format.
  • Webhook with URL and optional shared secret.

Each export can be active or inactive, use full_snapshot or incremental sync mode, and run on_change, on_demand, hourly, or daily. The export history table shows version, status, rows exported, sync mode, date, and error message.

Troubleshooting​

SymptomWhat to check
Master Data is not visibleConfirm the master_data.enabled feature is enabled and that a space is selected.
New Dataset is disabledThe space may have reached the master_data.max_datasets tier limit.
A dataset is visible in organization view but cannot be created thereCreation is intentionally space-scoped; switch into the target space.
Data tab says no schema is definedAdd columns on the Overview tab before entering rows.
Import succeeds with warningsReview unmapped headers, missing required columns, and type coercion warnings; only valid mapped row data is inserted.
Publish is blockedCheck foreign-key validation mode, required governance permissions, and whether the dataset is approval-required.
Export has no historyPublish a version or trigger active exports after configuring destinations.

Next Steps​