Currently, I am in the process of building a Node.js API that utilizes Express. As part of this project, I am incorporating the node-rest-client module to handle HTTP requests.
One of the key API endpoints that needs to be developed is /api/v1/users/:userId, which will provide comprehensive user information including details about the departments they are associated with.
In order to retrieve this information, the backend REST services required are:
/users/:userId - This service returns a JSON object containing user information as well as a list of department IDs. For example:
{ "name" : "xxx",
"departments" : [1, 5 ,6, 8]
}
/departments/:departmentId - This service provides a JSON object with department details:
{
"id" : x,
"name" : "xxx"
}
When making a call to /api/v1/users/1, the following steps need to be executed:
GET /user/1->{ "name" : "user1" , "departments" : [1, ,5 ,7 ,8]}- Retrieve the department IDs and make individual calls to
/departments/deparmentId - Once all relevant data has been collected, assemble the complete JSON response and return it.
I intend to optimize the request handling by parallelizing the operations using RxJs. I believe that utilizing Rx.Observable.zip() should suffice for this purpose.
The challenge lies in determining how to use Observable.zip() when dealing with an unknown number of Observables representing each HTTP request call within an array.
If the size of the array was fixed, calling zip() would look something like this:
var observables = [ obs1, obs2 ];
Rx.Observable.zip( observables[0], observable[1], function(...){...});
However, since the number of observables is variable, I am unsure of how to implement zip().