Tables

The first step to give you visibility on your data is to look into your tables. Tables are organised with the following hierarchy in the left menu:

1. Warehouse technology

2. Database

3. Schema

Each table is documented by a Detail Panel and 7 tabs:

• Tab Columns

• Tab Lineage

• Tab Read Me

• Tab Queries

• Tab Comments

• Tab History

• Tab Data Quality

Right Panel: Details

At the right, the Details menu summarises key attributes of the selected table. It is visible from all tabs

• Certified: Admins can mark a table as certified to indicate documentation has been reviewed ✅

• Deprecated: Admins can mark a table as deprecated to prevent users from using it ⚠️

• PII: Specifies if the table contains a column with Personal Identifiable Information

• Favorites: users can mark tables as favorites, it will appear on the homepage

• Owners: a user can be assigned as the owner of a table

• Popularity: computed based on the number of table read (select) queries in the last 30 days (not writes). Non human accounts are excluded of the count

• Latest data update: when turned on, Catalog displays when the table data was last updated

• Description: contains an extract of the description in the Read me tab. By clicking on it, you will return to the Read me tab.

• Completion level: computes how well the table has been documented

• Tags: tags can be used to describe tables. They are common for all users of a company.

• Frequent users: lists the power users of the table

• Table type: can either be a table, view, or an external table

• SQL source: shows the queries which create and update the selected table

Completion score is calculated in the following way:

• 20% of the score is attributed if a table owner is assigned

• 10% of the score is attributed if table Read me is complete

• 70% of the score is attributed if column description are complete

Links

Source Link

A Table has an URL of its own, to discover it in its own warehouse (Snowflake & BigQuery).

External Links

Associate different URLs with a Table, from the following applications:

• Github

• Gitlab

• Airflow

• Other

These links can be pushed with Catalog API.

There can currently exist one link of each type for one given Table.

Tab Columns

The Home tab lists all the columns present in the selected table and tags the related assets

Columns

Each column is described by 5 fields

Type

Name

Attribute

Description

The column type is specified by Catalog

Ex. Timestamp, integer, float, etc.

Column name

A column can be tagged as:

• containing Personal Identifiable Information (PII)

• being a primary key (PK)

There are several options to fill the description of each column:

• User can manually input a description

  • Auto complete: if you have already filled out the description of the same column in another table, Castor will suggest it to you

• Pre-filed documentation: Castor will centralise all the documentation you have already completed, in your warehouse or dbt as in the example below. The imported description will be tagged with the logo of its source.

Related Assets

By clicking on the Related assets button at the bottom of the page, a menu will appear to give you an overview of how the table is created and used.

Tab Lineage

You have three ways to view lineage

List View

Table Lineage

Column Lineage

In here you can view

• Source Tables: the most upstream Tables Catalog could find

• Parents: the direct upstream tables of your table

• Children: the direct downstream tables and dashboards

You can see there asset-level lineage.

• You can filter out in/out Dashboards.

• Also, since lineage is computed based on the observed queries from your warehouse, you can decide to only consider the queries from the last N days.

After picking the column you're interested in, you can see its own specific lineage.

Tab Read me

The 3rd tab of a Table's page contains the table's Read me

The Read me contains the selected table's description.

User inputed description

External description

• The first section of the Read me tab give space for users to input all relevant table descriptions and any

• This section is a Rich Text, allowing for maximum flexibility and formatting

• To view Rich Text options, simply highlight text portions

• Castor will centralise all the documentation you have already completed. The imported descriptions will be tagged with the logo of its source and will be immutable. Classic sources would be dbt or column comments in your warehouse

Pinned Assets

To improve your tables read-mes, you can directly pin relevant assets to it

How does this work? Checkout this page that details how it works

Tab Queries

The Queries tabs lists queries performed on the selected table

To find a query of interest, you can filter the queries by:

• Author: who wrote the query

• Joined tables: tables joined

To view the full query, simply click on it and a modal with appear:

To easily copy a query click either:

• Normal view: the top right icon of the query block

• Single query view: the icon by the author's name

No queries

If there is no query, it might be due to one of the following reasons:

• Query Age: The queries in question are older than 30 days.

• Excluded Sources: These are read queries generated by BI tools or other service accounts that have been intentionally excluded.

• Relevance: Some queries have been flagged as irrelevant, either because they are too simple or exceptionally large in scope.

Comments

To easily collaborate and communicate between users, each table has a Comments section.

History

A log of changes is kept under the History tab:

Data Quality

See data quality when browsing tables, along with a detailed breakdown of the tests and statuses.