HomeLogin

SQL Transformations

Introduction

SQL Transformations give you the full power of ANSI SQL to transform your events. Any transformation you can express with SQL, you can perform with SQL transformations. When you write a SQL transformation you will have access to an incoming_event table. This table has a number of columns that correspond to the built-in data Freshpaint collects and you will have access to an additional JSON column called props where you will have access to any custom properties. The full schema of the incoming_event table is described here.

SQL Transformations are part of the Freshpaint SQL add-on. Contact support@freshpaint.io to request access.

SQL Transformations are only available when either the event being transformed is a server-side event or the destination is in server-side mode. If you want to make as much use of SQL Transformations as possible, we recommend configuring your destinations in server-side mode.

Behavior

When you write a SQL Transformation you are writing a SQL query against the incoming_event table. Your query should return either a single row or no rows. Any columns returned by your query will be attached to the event as custom properties. If your query returns zero rows, the event will filtered out and not sent to any destinations. As an example, the query:

SELECT 10 AS example_prop FROM incoming_event;

Will create a new property example_prop with the value 10. If you return NULL for the value of a property, that property will instead be deleted.

There are two reserved column names that have special behavior when you return them:

  • event_name - When you return this from a query, the name of the event will become the provided value.

  • user_id - When you return this from a query, the user_id of the event will become the provided value.

Examples

Prepend a string to your events

This transformation will prepend all events with the prefix SQL: . If the incoming event name is Login, the transformed event will have the name SQL: Login.

SELECT CONCAT('SQL: ', event_name) AS event_name FROM incoming_event;

Run a custom formula

This example will double the value of the custom property a:

SELECT 2 * JSON_EXTRACT_DOUBLE(props, 'a') AS a FROM incoming_event;

Perform a regular expression match

This example extracts a URL parameter from the url of the event using regular expressions:

SELECT REGEXP_REPLACE(url, '^.*param=(\\w+).*$', '\\1') AS param FROM incoming_event;

Add a constant property

This example creates a new property called new_property an always sets it to the value 10:

SELECT 10 AS new_property FROM incoming_event;

Delete a property

This example deletes the property sensitive_identifier from the event. This is done by setting the value of the property to NULL.

SELECT NULL AS sensitive_identifier FROM incoming_event;

Filter events based on a regular expression

This example filters out all events where the email is a gmail address:

SELECT event_name 
FROM incoming_event 
WHERE JSON_EXTRACT_STRING(props, 'email') NOT RLIKE 'gmail.com$';

Schema

The incoming_event table has the following schema:

PreviousModify User DataNextInvestigate: Testing and Debugging

Last updated