Collibra Catalog Custom Fields

Overview

The Collibra - Catalog Custom Fields set adds a Collibra section to the Validatar catalog at data source, schema, table and column level. These fields hold what Collibra knows about each object: its asset id, full name and link, description, status, owners and stewards, tags, business terms, data categories, security classification and personal data flags, and the data classes Collibra has accepted for each column.

The Collibra metadata script fills the fields from Collibra. The Collibra Data Source Template reads two of them when it publishes Validatar results to Collibra. Both are part of the Collibra Integration.

Marketplace: Collibra - Catalog Custom Fields

Collibra descriptions, tags and owners land in their own fields. Validatar's own Description, Tags and Owner are never touched, so nothing your team maintains in Validatar is overwritten.

Note: These are Validatar fields: they hold Collibra's metadata inside Validatar. Results travel the other way as Data Quality Rules in Collibra, which Collibra Setup for Validatar covers.

Fields

Level	Field	Reference key	Type	What it holds
Data source	Collibra Database Full Name	`collibra_database_full_name`	Short text	The database's full name in Collibra, such as `Snowflake-Connection>SAMPLE_DW`: the prefix of every asset under it
Data source	Collibra Database Name	`collibra_database_name`	Short text	The database's name in Collibra
Data source	Collibra Edge Connection	`collibra_edge_connection`	Short text	The Edge connection the database was registered through
Data source	Collibra Database Domain	`collibra_database_domain`	Short text	The domain that holds the database's assets
Data source	Collibra Database Link	`collibra_database_url`	Short text	Link to the database in Collibra
Data source	Collibra Data Source Synced	`collibra_data_source_synced`	Date	When the metadata script last updated the data source
Schema	Collibra Schema Full Name	`collibra_schema_full_name`	Short text	Collibra's full name for the schema
Schema	Collibra Schema Link	`collibra_schema_url`	Short text	Link to the schema in Collibra
Schema	Collibra Schema Description	`collibra_schema_description`	Paragraph	The schema's description in Collibra
Schema	Collibra Schema Status	`collibra_schema_status`	Short text	The schema's status in Collibra
Schema	Collibra Schema Tags	`collibra_schema_tags`	Tag list	Collibra tags on the schema
Schema	Collibra Schema Domain	`collibra_schema_domain`	Short text	The domain that holds the schema
Schema	Collibra Schema Synced	`collibra_schema_synced`	Date	When the metadata script last updated the schema
Table	Collibra Asset ID	`collibra_asset_id`	Short text	The table's asset id in Collibra. Publishing uses it to find the table
Table	Collibra Full Name	`collibra_full_name`	Short text	The table's full name, such as `Snowflake-Connection>SAMPLE_DW>STAR_SCHEMA>FACT_SALES`
Table	Collibra Link	`collibra_url`	Short text	Link to the table in Collibra
Table	Collibra Description	`collibra_description`	Paragraph	The table's description in Collibra
Table	Collibra Status	`collibra_status`	Short text	The table's status, such as Accepted, Obsolete or Missing from source
Table	Collibra Owners	`collibra_owners`	Short text	People with the Owner role on the table, as names and emails
Table	Collibra Business Stewards	`collibra_business_stewards`	Short text	People with the Business Steward role
Table	Collibra Technical Stewards	`collibra_technical_stewards`	Short text	People with the Technical Steward role
Table	Collibra Tags	`collibra_tags`	Tag list	Collibra tags on the table
Table	Collibra Business Terms	`collibra_business_terms`	Tag list	Business terms that represent the table
Table	Collibra Data Categories	`collibra_data_categories`	Tag list	Data categories the table belongs to
Table	Collibra Security Classification	`collibra_security_classification`	Short text	The table's security classification in Collibra
Table	Collibra PII	`collibra_pii`	Dropdown: Yes, No	Whether Collibra marks the table as holding personally identifiable information
Table	Collibra Domain	`collibra_domain`	Short text	The domain that holds the table
Table	Collibra Community	`collibra_community`	Short text	The community of that domain
Table	Collibra Last Modified	`collibra_last_modified`	Date	When the table's asset last changed in Collibra
Table	Collibra Last Synced	`collibra_last_synced`	Date	When the metadata script last updated the table
Table	Publish Scores to Collibra	`collibra_publish_scores`	Dropdown: Yes, No	No keeps this table out of Collibra when Validatar publishes. Blank or Yes publishes it
Column	Collibra Column Asset ID	`collibra_column_asset_id`	Short text	The column's asset id in Collibra
Column	Collibra Column Link	`collibra_column_url`	Short text	Link to the column in Collibra
Column	Collibra Column Description	`collibra_column_description`	Paragraph	The column's description in Collibra
Column	Collibra Column Status	`collibra_column_status`	Short text	The column's status in Collibra
Column	Collibra Column Tags	`collibra_column_tags`	Tag list	Collibra tags on the column
Column	Collibra Column Business Terms	`collibra_column_terms`	Tag list	Business terms that represent the column
Column	Collibra Column Data Classes	`collibra_column_data_classes`	Tag list	Data classes accepted for the column by Collibra's data classification
Column	Collibra Column PII	`collibra_column_pii`	Dropdown: Yes, No	Whether Collibra marks the column as personally identifiable information
Column	Collibra Column Synced	`collibra_column_synced`	Date	When the metadata script last updated the column

That is 40 fields: 6 at data source level, 7 at schema level, 18 at table level and 9 at column level. Table fields have the short names, and the fields at every other level name their level. Validatar requires a custom field's name and reference key to be unique across all catalog levels, so the schema's link is collibra_schema_url rather than a second collibra_url.

What fills each field

The Collibra metadata script fills every field except Publish Scores to Collibra. You choose which fields it sends, and whether it covers columns, at the top of the script.
People come in as text, such as Dana Steward <dana@acme.example>, so a Collibra user who is not a Validatar user never fails the run.
Statuses and security classifications are text fields, because every Collibra instance defines its own values. A Dropdown would reject the whole run on a value it does not list.
Collibra Column Data Classes holds only the classifications a steward accepted. Suggestions nobody has reviewed are left out; the health check Data class suggestions reviewed lists them.
Publish Scores to Collibra is set by your team, not by the script. Set it to No on a table to keep that table out of Collibra when Validatar publishes.
Collibra Asset ID on a table is how the publish macro finds the table in Collibra when it is filled. Without it, the macro matches schema and table names under the Collibra database that the template's connection map names.

The Collibra section on a table in the Validatar catalog

Import the fields

You need Govern Edition and the Global Custom Metadata Admin role.

From the marketplace (recommended)

In Validatar, open Marketplace and find Collibra - Catalog Custom Fields.
Choose Get, then Import Directly. Validatar opens Settings > Catalog Custom Fields with an Import Custom Fields panel on the right. The panel's list holds the Collibra section for each level: Data Source, Schema, Table and Column.
For each level, open that level on the page, pick the same level in the panel's list, choose Import All once, then Save.
Check the Data Source, Schema, Table and Column levels. Each should now have a Collibra section with the fields listed above.

Warning: Choose Import All only once per level. A second Import All adds the same fields again, and Save then refuses with "Please fix the errors on this page" without marking a field. If that happens, leave the page without saving, reload it and import that level again.

From the downloaded file

If a level has no Collibra section after the marketplace import, or you are moving the set between Validatar instances:

Choose Get, then Download XML.
In Settings > Catalog Custom Fields, open the level, choose Import, select the file and choose Next.
Choose Import All, then Save. Repeat for each level.

Note: Import before the first run of the metadata script. Validatar ignores a script column whose field does not exist, so the run succeeds but the value is not stored.

Test-level publishing fields

Publishing to Collibra also reads five fields on each test. They are project custom fields at Test level, so they are not part of this set. The Collibra Project Custom Fields item adds them, in a section named Collibra Publishing.

Add your own Collibra attributes

To bring another Collibra attribute across, such as a retention period, add a table-level field in the Collibra section with a reference key that no other level uses, for example collibra_retention. Then name it in the metadata script as "Collibra attribute type name": "reference key":

COLLIBRA_ATTRIBUTES = {
    "Data Retention Period": "collibra_retention",
}

Use Short text unless you know every value Collibra can send. A Dropdown field rejects the whole run when Collibra sends a value that is not one of its options.

Note: The field set has been imported from the Validatar marketplace into a Validatar instance and filled there by the Collibra metadata script from a built-in sample Collibra instance: 6 data source, 12 schema, 102 table and 78 column values on a Snowflake data source. It has not yet been used against a live Collibra instance, so expect small adjustments after your first live run.