Data Validation in Scripted REST API

[greg54]

Hi,

For the people who are using Scripted REST API's to post and update data to ServiceNow, how do you handle data validation? Specifically for reference fields? I have found that if you pass in an invalid sys_id, the record is still created with a null value where the sys_id was invalid. This is unexpected behavior, as I would expect an error response of an invalid value.

Am I missing something trivial here or is the developer responsible for handling all data validation?

Thanks!

[Nathan Grove]

I always manually look up the referenced field for example:

Say I'm inserting an incident with a task referenced on it, I will manually check that the task exists before doing the insert. The task sys id is passed in on the request body in the field "task".

// validate the task sys_id
var task = new GlideRecord('task');

// get the task by sys id...if the fetch fails...then we know it doesnt exist...

if (! task.get(body.task)){

  // some HTTP code to indicate something is wrong

  response.setStatus(400);

}

// build and insert the incident

[greg54]

Thanks for the reply Nathan. Good to know. Wanted to be sure that was the only way before going down that path.

[robinson_gregor]

One more quick question. What is the difference between the functionality of a Scripted REST API and a Script type Processor? Both seem to do exactly the same thing from what I have researched.

Is there anything I am missing where one has an advantage of the other?

Thanks!

[Chuck Tomasi]

Scripted REST API should be used over processors. They have the following advantages:

• Better control over the request and response (and parameters)
• Versioning (this is the big one). You can leave your remote clients using v1 while you enhance v2 and have both deployed at the same time!

This should help

Scripted REST APIs

[robinson_gregor]

Thank you so much Chuck.