The short of it is…Power Platform lives on Azure. Now this article is not an attempt to push people into Azure, but I do at least want you to consider it. If you are like me, your introduction into Azure was most likely Azure AD now known as Entra ID and next year probably something else.
Azure is a big universe that supports many different workloads. Of particular interest to me and for the benefit of this article I want to focus on Azure Functions.
What Are Azure Functions?
I don’t think there is a need to dive deep into the subject, but what I will say is, Azure Functions allow us to run serverless, on-demand code. Azure Functions support various programming languages such as C# and JavaScript. The ability to call these functions and have them execute a piece of code is beneficial to developers, especially to “low-coders”. Reason being is, we don’t have to be as seasoned as a traditional “pro-coder” we can focus only on the code that matters to us.
With the advent of Co-Pilot and ChatGpt we can write code that serves our needs. I’d hate to cause a stir so I will say this, be careful with a.i. generated code, make sure your prompt is as explicit as you can get it and if you are unsure about the piece of code you are using…ask for help.
Exercise
I once worked with a client who needed to compare two objects and identify their differences to create exceptions for their customer service representatives to address. Implementing this process in Power Automate significantly saved time, though the solution involved multiple actions, loops, and data transformations, making it less streamlined than desired. However, as I gained more expertise in Azure Functions, I discovered that the same task could be accomplished more efficiently using just a few lines of JavaScript and a single action within the flow.
Before we proceed with this exercise, it’s important to note that a foundational understanding of how to provision an Azure resource and set up your function app is essential. To ensure you’re adequately prepared, I strongly recommend reviewing the information provided in the following link: Getting started with Azure Functions | Microsoft Learn.
Keep in mind that we are utilizing JavaScript for this task, so make sure to configure your function accordingly. While you can create your function directly in the Azure portal, which is the simplest approach, there’s also the option to develop it in Visual Studio Code. This latter method is a bit more complex but certainly achievable if you’re up for the challenge. For our function app, we’ll be using an HTTP trigger as the primary mechanism to initiate the process.
Code
Keep in mind that a function app essentially serves as a container for your code snippet. In this exercise, we’ll be adding code to our function and then obtaining the URL to integrate it with Power Automate. While I utilized ChatGPT to generate my code, it’s crucial to be aware that certain parts of this code are indispensable. One such essential line of code is as follows:
module.exports = async function (context, req){}It’s important to operate within this framework, as the context from Power Automate will be fed into this function. Once that’s in place, our code will handle the rest of the process. This setup ensures that the data from Power Automate is seamlessly integrated and processed by our Azure Function.
Full code
Simply copy and paste the entire code into your function. Pay special attention to the parameter names ‘object1’ and ‘object2’. These are the parameters through which we will pass our objects. The code will then perform the comparison and return an object highlighting the differences, which we can utilize on the Power Automate side.
module.exports = async function (context, req) {
context.log('JavaScript HTTP trigger function processed a request.');
try {
// Assuming the input is a JSON string representing the objects.
const obj1 = req.query.object1 ? JSON.parse(req.query.object1) : null;
const obj2 = req.query.object2 ? JSON.parse(req.query.object2) : null;
// Validate the input
if (typeof obj1 !== 'object' || obj1 === null || typeof obj2 !== 'object' || obj2 === null) {
context.res = {
status: 400,
body: "Please pass two JSON objects in the query string or in the request body",
headers: {
'Content-Type': 'application/json; charset=utf-8'
}
};
return;
}
// Find differences
const differences = findObjectDifferences(obj1, obj2, "root");
context.log('Differences:', differences);
// Return the differences as a JSON string with the header "differences"
context.res = {
body: { "differences": differences },
headers: {
'Content-Type': 'application/json; charset=utf-8'
}
};
} catch (error) {
context.res = {
status: 500,
body: "Error processing your request",
headers: {
'Content-Type': 'application/json; charset=utf-8'
}
};
context.log.error("Exception thrown", error);
}
};
function findObjectDifferences(obj1, obj2, path) {
let differences = [];
let allKeys = new Set([...Object.keys(obj1), ...Object.keys(obj2)]);
allKeys.forEach((key) => {
let newPath = path + "." + key;
if (obj1.hasOwnProperty(key) && !obj2.hasOwnProperty(key)) {
differences.push({ path: newPath, obj1: JSON.stringify(obj1[key]), obj2: 'Not present' });
} else if (!obj1.hasOwnProperty(key) && obj2.hasOwnProperty(key)) {
differences.push({ path: newPath, obj1: 'Not present', obj2: JSON.stringify(obj2[key]) });
} else if (obj1[key] && obj2[key] && typeof obj1[key] === 'object' && typeof obj2[key] === 'object') {
differences = differences.concat(findObjectDifferences(obj1[key], obj2[key], newPath));
} else if (obj1[key] !== obj2[key]) {
differences.push({ path: newPath, obj1: JSON.stringify(obj1[key]), obj2: JSON.stringify(obj2[key]) });
}
});
return differences;
}
After configuring your function, save it. Then, locate and click on the ‘Get function URL’ button found in the editor’s ribbon. This URL is crucial as you will need it to make the HTTP request in Power Automate.
Power Automate
- Set Up the Trigger: Begin by setting up your Power Automate flow with a trigger of your choice. This trigger will initiate the flow.
- Initialize Object Variables: Next, initialize two object variables named ‘varObject1’ and ‘varObject2’. Initially, I attempted to directly write the objects in the HTTP request but encountered issues. It’s more effective to initialize these variables and assign your arrays to them, allowing for a smoother conversion process.
- Create an HTTP Request:
- Choose ‘POST’ as the method for your HTTP request.
- In the query, you will need to add two lines:
- The first line should have ‘object1’ as the key, with the value being the ‘varObject1’ variable.
- The second line should have ‘object2’ as the key, with the value being the ‘varObject2’ variable.
- Paste the Function App URL: Paste the URL you obtained from your Azure Function app into the appropriate field in your HTTP request setup.
- Save and Run: After setting up everything, save your flow configuration. Then, run your flow to test and see it in action.
The following is the returned exceptions (differences) in the objects. The object specifies the affected columns “root.Borrower2” and “root.Address”
{ "differences": [ { "path": "root.Borrower2", "obj1": "\"Marsurice DeVan\"", "obj2": "\"Marsurice Devan\"" }, { "path": "root.Address", "obj1": "\"Dean\"", "obj2": "\"Down\"" } ] }