This section explains the conventions used when building the Snowflake database and functionality for implementing various features such as history tracking and support for ETL.
Supported features
Data update methodology
Gong data is made available in data warehouses using different methodologies:
Snowflake: get your data via a data share
Databricks and BigQuery: changed data is transferred daily
ETL support
Each table includes ETL_MODIFIED_DATETIME column which records the date and time the entity was last updated.
History tracking
History is tracked in the USERS table using Slowly Changing Dimension (SCD) type 2.0 with the VALID_FROM_DATETIME and VALID_TO_DATETIME columns. The VALID_TO_DATETIME is NULL for data that is currently valid. Historical data is available from August 2022. Contact us if you need earlier data.
Update frequency
The Gong database is updated a maximum of 24 hours after the data is available in Gong.
Soft delete
Certain tables contain the is_deleted field to support soft delete of data. The is_deleted field by default is false, and is changed to true for records that were deleted in Gong. When a record is deleted in Gong, some of the data for that record is deleted from Snowflake as well. The record's primary keys are retained.
Schema and naming conventions
This is a list of conventions used when designing the database:
The schema is in 3rd normal form. Some data may be denormalized for query ergonomics.
Table names are plural and snake_case.
Column names are SNAKE_CASE.
Id columns are suffixed with _id and are of varchar type and should be treated as such regardless of their content.
Date and time column names are suffixed with _datetime and are of type TIMESTAMP_TZ which includes UTC offset.
Unique ID column
Each table contains a ROW_ID column which acts as the unique identifier for each row in the table.
Date format
All dates are in the timestamp_tz format in the UTC timezone.