Collections
Collections store Functions, Tables, and Connections.
Collection Purpose, Behaviour and Architecture varies depending on whether they store Publisher, Transformers, or Subscribers
Transformer Collections
Transformer collections store Transformer functions, which read input data from Tabsdata Tables and write output data into other Tabsdata Tables.
Transformer Collections can store one or more Tabsdata Functions. Since no calls to external systems are made through these collections, they do not require a Connection to be bound to it.
The main purpose of Transformer collections is that they define commit behavior during Execution Plans. For more information, view [Pristine State] and [Transactions]
Publisher Collections
Publisher Collections store Publisher Functions, which read input data from external systems and write output data into Tabsdata Tables.
While Publishers connect to external systems, they do not contain any credentials to those systems, which are stored in a separate object called a connection
The main purpose of Publisher Collections is to bind a Publisher to a Connection, allowing the Function to use the credentials stored in that Connection
Unlike Transformer Collections, Publisher Collections can only hold one Publisher at a time and must also have a Connection bound to it.
Subscriber Collections
Subscriber Collections store Subscriber Functions, which read input data from Tabsdata Tables and write output data to external systems.
While Subscribers connect to external systems, they do not contain any credentials for those systems. Instead, credentials are stored in a separate object called a Connection.
The main purpose of Subscriber Collections is to bind a Subscriber to a Connection, allowing the Function to use the credentials stored in that Connection.
Unlike Transformer Collections, Subscriber Collections can only hold one Subscriber at a time and must also have a Connection bound to them.