Working with Triggers

Triggers execute Tabsdata functions. A trigger can be initiated manually through a CLI command or automatically by a new commit to its associated table.

Manual Triggers

Manual triggers are initiated by a CLI command. For example, you can use a CLI command to manually trigger a publisher to read sales data from a database table and publish it to the Tabsdata server.

You can manually trigger any Tabsdata function, even those with a specified automated trigger.

Using Manual Triggers

To trigger a function, use the following command in a command line interface:

$ td function trigger --collection <collection_name> --name <function_name>

Where:

• <collection_name> is the name of the collection in the Tabsdata server that the function is registered with.

• <function_name> is the name of the function to be executed. This name is defined when the function is registered.

Automated Triggers

Automated triggers are initiated by a new commit to one of the tables associated with the function. For example, you might define an inventory subscriber function that exports data from a Tabsdata inventory table to Amazon S3 every time the inventory table receives a commit.

Note that a commit does not always include changes in the table data. For example, a publisher that writes to three tables performs a commit to all three tables with each execution, irrespective of whether there is a change in data for any of the tables.

Since, changes on tables can automatically trigger functions which in turn create a new commit to other tables, Tabsdata enables a cascading workflow of updates to downstream tables when the uptream data changes.

trigger_by

Triggers in Tabsdata functions are defined using the trigger_by property.

Below are the different ways in which trigger_by can be configured:

Not Defining

If the trigger_by property is not defined, the behavior depends on the default behaviour of the function:

Publishers: By default, publishers do not have any triggers. If trigger_by is not defined, the publisher can only be triggered manually.

Transformers: By default, transformers are triggered by a commit to any of their input tables. Hence, if trigger_by is not defined, the transformer will automatically respond to new commits to any of its input tables. To specify different trigger tables, define the trigger_by tables using a non-empty array.

Subscribers: By default, subscribers are triggered by a commit to any of its input tables. Hence, if trigger_by is not defined, the subscriber will automatically respond to new commits to any of its input tables. To specify different trigger tables, define the trigger_by tables using a non-empty array.

Defining an Empty Array

trigger_by = []

Setting trigger_by to an empty array disables automatic triggering. This is useful for transformers and subscribers when you want to prevent them from being triggered by their input tables.

Defining a Non-Empty Array

trigger_by = ["<table_name>", "<collection_name>/<table_name>"]

Setting trigger_by to a non-empty array explicitly defines which tables can trigger the function. If you want the functions to be triggered by a subset of their input tables or by external tables, list them in the array. To reference tables outside the function’s collection, use the format <collection_name>/<table_name>.

Only those tables listed in the trigger_by array will trigger the function. List as many as needed.