Data Operations — Compose, Parse JSON, Filter Array

Learn how to manipulate and transform data within your Logic Apps workflows using built-in data operation actions.

Prerequisites

• An Azure account with an active subscription
• Basic familiarity with Logic Apps (see previous tutorial)
• A Logic App (Consumption or Standard) created in the Azure portal

Sample Data

We'll use this sample order array throughout the tutorial:

[
  { "id": 1, "product": "Laptop", "price": 1200, "category": "Electronics", "inStock": true },
  { "id": 2, "product": "Desk Chair", "price": 350, "category": "Furniture", "inStock": true },
  { "id": 3, "product": "Monitor", "price": 800, "category": "Electronics", "inStock": false },
  { "id": 4, "product": "Notebook", "price": 15, "category": "Stationery", "inStock": true },
  { "id": 5, "product": "Keyboard", "price": 150, "category": "Electronics", "inStock": true }
]

1. Compose Action

The Compose action lets you create JSON objects, merge data, or construct any output without needing an external service.

Use Cases

• Build a JSON payload to pass to a later action
• Merge values from multiple sources into a single object
• Store intermediate calculations

Step-by-Step

1. In your Logic App designer, click + New step.
2. Search for Compose under Data Operations.
3. In the Inputs field, enter your JSON object or expression.

Example — Create a JSON object:

{
  "orderSummary": {
    "totalItems": 5,
    "generatedAt": "@{utcNow()}",
    "source": "AzureIntegrationHub"
  }
}

Example — Merge dynamic content:

{
  "fullName": "@{triggerBody()?['firstName']} @{triggerBody()?['lastName']}",
  "email": "@{triggerBody()?['email']}"
}

Tip: Use the Compose action output in subsequent steps by referencing outputs('Compose').

2. Parse JSON Action

The Parse JSON action takes a JSON string or object and produces typed dynamic content you can use in later steps.

Step-by-Step

1. Add a Parse JSON action (under Data Operations).
2. In Content, select the output from a previous step (e.g., an HTTP response body).
3. In Schema, provide the JSON schema.

Generating a JSON Schema from a Sample Payload

1. Click Use sample payload to generate schema.
2. Paste your sample JSON:

{
  "id": 1,
  "product": "Laptop",
  "price": 1200,
  "category": "Electronics",
  "inStock": true
}

3. Click Done. The schema is auto-generated:

{
  "type": "object",
  "properties": {
    "id": { "type": "integer" },
    "product": { "type": "string" },
    "price": { "type": "integer" },
    "category": { "type": "string" },
    "inStock": { "type": "boolean" }
  }
}

Using Dynamic Content After Parsing

Once parsed, each property (e.g., product, price) appears as selectable dynamic content in subsequent actions.

Tip: If your payload may have optional fields, set "required": [] in the schema or mark fields as nullable to avoid runtime failures.

3. Filter Array Action

The Filter Array action returns only the items from an array that match a specified condition.

Step-by-Step

1. Add a Filter array action.
2. In From, provide the source array (e.g., output of a Parse JSON or Compose action).
3. Set the condition using the Advanced mode expression or the visual editor.

Example — Filter electronics that are in stock:

• From: @body('Parse_JSON')
• Condition: @item()?['category'] is equal to Electronics AND @item()?['inStock'] is equal to true

Result:

[
  { "id": 1, "product": "Laptop", "price": 1200, "category": "Electronics", "inStock": true },
  { "id": 5, "product": "Keyboard", "price": 150, "category": "Electronics", "inStock": true }
]

4. Select Action

The Select action transforms each item in an array into a new shape (like a map function).

Step-by-Step

1. Add a Select action.
2. In From, provide the source array.
3. In Map, define the output object shape using key-value pairs.

Example — Extract product name and price only:

• From: @body('Filter_array')
• Map:

Result:

[
  { "ProductName": "Laptop", "Price": 1200 },
  { "ProductName": "Keyboard", "Price": 150 }
]

5. Create CSV Table / Create HTML Table Actions

These actions convert an array of objects into a CSV string or an HTML table.

Create CSV Table

1. Add a Create CSV table action.
2. In From, provide the array (e.g., output of the Select action).
3. Choose Automatic columns (uses all properties) or Custom to pick specific columns.

Output:

ProductName,Price
Laptop,1200
Keyboard,150

Create HTML Table

1. Add a Create HTML table action.
2. In From, provide the array.
3. Choose column mode.

Output:

<table>
  <thead><tr><th>ProductName</th><th>Price</th></tr></thead>
  <tbody>
    <tr><td>Laptop</td><td>1200</td></tr>
    <tr><td>Keyboard</td><td>150</td></tr>
  </tbody>
</table>

Tip: Combine the HTML table output with the Office 365 connector to send formatted email reports.

Complete Workflow Example

Here's the full flow combining all actions:

1. Trigger — When an HTTP request is received (with the sample order array)
2. Parse JSON — Parse the request body with the order schema
3. Filter Array — Keep only Electronics items in stock
4. Select — Map to ProductName and Price
5. Create HTML Table — Generate a formatted table
6. Send Email — Send the HTML table in an email body

Tips for JSON Schema Generation

• Always use Use sample payload to generate schema rather than writing schemas manually.
• Test with multiple sample payloads to ensure all possible fields are captured.
• For arrays, provide at least one item that includes all fields.
• Add "required": [] if fields may be missing at runtime.
• Use "type": ["string", "null"] for nullable fields.

Common Issues

Next Steps

• Previous: Connectors — HTTP, Office 365, SQL Server
• Next: Variables and Loops