Finding Assets

Finding relevant data assets in your organization is one of the main use cases for Katalogue. There are two main ways to find assets - browsing and searching.

Browse

Assets in Katalogue are organized in two main hierarchial categories, or tree-like structures. These can be explored by hierarchical browsing, or drill-down navigation:

• Datasets. The data your organization actually holds, from systems and tables to files, APIs and fields. Can be thought of as data models. Contains assets in the following hierarchial order:
  1. Systems
  2. Datasources
  3. Dataset Groups
  4. Datasets
  5. Fields
• Glossaries. The words your organization uses, what they mean, and where the data they describe lives. Can be thought of as information models. Contains assets in the following hierarchial order:
  1. Glossaries
  2. Entities
  3. Attributes
  4. Attribute Values

Attributes are the point where the two meet: an attribute is a re-usable description that belongs to a glossary and can be associated with the fields that carry it, so a field inherits its description from the glossary. Since an attribute is useful to browse on its own, regardless of which glossary it sits in, the Glossaries page has an All Attributes tab listing every attribute in Katalogue.

Each level opens onto the next, so you do not have to walk down one page at a time. A dataset group lists its datasets, and a dataset opens to show its fields — including the fields nested inside other fields, which is how an API response describes its JSON. A glossary lists its entities and attributes, and a Dimension opens to show its values.

A glossary’s asset list, and the same list on an entity, can also be shown as a flat list: every entity and attribute at one level, sorted by name, with the assets it sits under shown beneath its name. It is the quicker way to answer “is there already a term for this?”, or to scan for the assets that still have no description, since sorting then orders the whole glossary rather than each parent’s children. Attribute Values still open under their attribute, and the choice is remembered between visits.

Examples

1. Select the “Datasets” option in the Katalogue navigation pane to list all systems.
2. Click a system in the list to see its details and a list of all datasources in the system.
3. Click a datasource in the list to see its details and a list of all dataset groups in the system.
4. Click the ”< BACK” button to go back to the last visited asset, i.e. go up one level in the hierarchy to the list of datasources.
5. Select another datasource to see its details and a list of all dataset groups in the system.
6. Etc until you find what you are looking for.
7. Expand a dataset to see the fields it holds without leaving the list.
8. Click a field to see more details about the field.

Search

Katalogue comes with a simple, yet powerful search function to find relevant assets.

Technically, the search functionality is based on the PostgreSQL Full Text Search feature, see the Architecture page for more details.

Examples

Simple Search

The default behaviour is wildcard match with “and”-logic.

This searches for data assets containing words that begin with “customer” and “name”:

customer name

Context Search

Use a colon to invoke the context search index. This can be used to limit your search to a certain context.

This searches for data assets matching “customer” in systems, datasets etc that match “crm”:

customer : crm

Exact Name Search

Use a space at the end to find assets with exact names.

This searches data asset names that exactly matches “customer_name”:

customer_name // with a space at the end

Phrase Search

Use quotation marks to search for exact phrase matches.

This searches for data assets matching exactly the word “customer” directly followed by “name”:

"customer name"

Or Logic

Use a pipe to match assets with “or” logic.

This searches for data assets containing words that begin with either “customer” or “name”:

customer | name

Exclude Words

Use an exclamation mark to exclude words.

This searches for data assets matching “customer” and that does not include the word “name”:

customer ! name