Data Transformsformation with JSON
Hello,
I would like to describe two REST service scenarios , and then ask a question.
Scenario 1)
- We invoke a REST service that returns XML .
- In order to map the data into the application class structure, an XML Parse rule is used.
- The parse xml rule maps the incoming data into the application class structure.
- Done.
Scenario 2)
- We invoke a REST service that returns JSON. PRPC automatically parses the JSON and populates a page in the clipboard.
- In order to map the data into the application class structure, a data transform rule is used.
- To use the data transform , we need to create classes within PRPC that exactly mirror the structure of the incoming data, otherwise the transform will not save.
- The data transform rule maps the incoming data into the application class structure.
- Done
QUESTION :
Why is step 3 of scenario 2 needed? That step is not an insignificant amount of work.
Message was edited by: Vidyaranjan Av | Included Category
Hi Joseph,
In general, data transformation involves both mapping data from a source to a target, and performing any conversions or other transformations on that data that might be required to achieve the intended mapped results .
Even though an activity could be used to set many property values on a page, using a data transform speeds development, because the data transform form is easier to understand than the Activity form.
Yes to parse the response we should have create classes which will match exact structure of incoming data .
Kindly let us know what do you mean exactly by that step is not an insignificant amount of work?
Rulesware LLC
CA
Hi Arun,
Creating the class structure for incoming data is NOT needed when the data is in XML format.
Why is loose coupling promoted in one technology (XML) and tight coupling promoted in another (JSON) ?
Shouldn't loose coupling between the model and view be promoted at all times, regardless of data source or technology ?
Thanks
Accepted Solution
Pegasystems Inc.
CA
Hi Joseph,
I understand your concern. We are continuously improving our mapping capabilities as far as JSON is concerned and you can expect further refinements in this area in future releases. Unfortunately, at this point, if you want to map JSON data directly to application class structure, you have two options. The issue stems from the fact that we don't have a Parse XML equivalent for JSON.
1. (Recommended) Create class structure corresponding to JSON data and use data transform/activity to map the data to application class structure. You haven't mentioned which version of Pega you are using. But in 7.x series, we have a REST wizard (that's primarily meant to create REST connectors) that you could leverage to upload your JSON sample and create its corresponding class structure in Pega. This is helpful in your case as you mentioned that creating the class structure manually isn't an insignificant amount of work.
2. (Not recommended) Even without class structure, as you observed, JSON data is mapped to a page on the clipboard. However, all these properties are virtual properties. So you need to write Java step in activity to read data from this page and map the data to your application class structure. I wouldn't recommend this though, especially when you have a large number of properties to map.
Rulesware LLC
CA
Thanks for the info Praneeth.
Rulesware LLC
CA
Thanks for the information Purighalla.
Since our needs cannot be addressed with inbuilt functionality, we will have to create our own custom solution. This thread of conversation was instrumental in helping us arrive at that conclusion.
Thanks again.