An Azure account and subscription. If you don't have a subscription, sign up for a free Azure account .

The logic app workflow where you want to perform the data operation. Both Consumption and Standard logic app workflows support the data operations described in this guide.

All data operations are available only as actions. So, before you can use these actions, your workflow must already start with a trigger as the first step and include any other actions required to create the outputs that you want to use in the data operation.

Data operation actions

The following sections summarize the data operations you can use. The operations are organized based on the source data types that the operations work on, but each description appears alphabetically.

JSON actions

The following actions help you work with data in JavaScript Object Notation (JSON) format.

To create more complex JSON transformations, see Perform advanced JSON transformations with Liquid templates .

Array actions

The following actions help you work with data in arrays.

Compose action

To construct a single output such as a JSON object from multiple inputs, use the action named Compose . Your inputs can have various types such as integers, Booleans, arrays, JSON objects, and any other native type that Azure Logic Apps supports, for example, binary and XML. You can then use the output in actions that follow after the Compose action. The Compose action also helps you avoid repeatedly entering the same inputs while you build your logic app's workflow.

For example, you can construct a JSON message from multiple variables, such as string variables that store people's first names and last names, and an integer variable that stores people's ages. In this example, the Compose action accepts the following inputs:

{ "age": <ageVar>, "fullName": "<lastNameVar>, <firstNameVar>" }

And creates the following output:

{"age":35,"fullName":"Owens,Sophia"}

To try the Compose action, follow these steps by using the workflow designer. Or, if you prefer working in the code view editor, you can copy the example Compose and Initialize variable action definitions from this guide into your own logic app's underlying workflow definition: Data operation code examples - Compose . For more information about the Compose action in the underlying JSON workflow definition, see the Compose action .

Follow these general steps to add the Data Operations action named Compose .

On the designer, select the Compose action, if not already selected. In the Inputs box, enter the inputs to use for creating the output.

For this example, follow these steps:

In the Inputs box, enter the following sample JSON object, including the spacing as shown:

In the JSON object, put your cursor in the corresponding locations, select the dynamic content list (lightning icon), and then select the corresponding variable from the list:

Follow these general steps to add the Data Operations action named Compose .

On the designer, select the Compose action, if not already selected. In the Inputs box, enter the inputs to use for creating the output.

For this example, follow these steps:

In the Inputs box, enter the following sample JSON object syntax, including the spacing as shown:

For each variable declaration in the object syntax, place the insertion cursor as described in the following table. Select the dynamic content list (lightning icon), and then select the corresponding definition from the Variables list:

Test your workflow

To confirm whether the Compose action creates the expected results, send yourself a notification that includes output from the Compose action.

In your workflow, add an action that can send you the results from the Compose action.

This example continues by using the Office 365 Outlook action named Send an email (V2) :

In this action, for each box where you want the results to appear, select inside each box, and then select the dynamic content list (lightning icon). In the list, under the Compose action, select Outputs .

For this example, the result appears in the email body, so you add the Outputs field to the Body box:

Save your workflow, and then manually run your workflow. On the designer toolbar, select Run > Run .

If you use the Office 365 Outlook - Send an email action, the following example shows the result:

Create CSV table action

To create a comma-separated value (CSV) table that has the properties and values from JavaScript Object Notation (JSON) objects in an array, use the action named Create CSV table . You can then use the resulting table in actions that follow the Create CSV table action.

To try the Create CSV table action, follow these steps by using the workflow designer. Or, if you prefer working in the code view editor, you can copy the example Create CSV table and Initialize variable action definitions from this guide into your own logic app's underlying workflow definition: Data operation code examples - Create CSV table . For more information about the Create CSV table action in the underlying JSON workflow definition, see the Table action .

Follow these general steps to add the Data Operations action named Create CSV table .

On the designer, select the Create CSV table action, if not already selected. In the From box, enter the array or expression to use for creating the table.

For this example, select inside the From box, and select the dynamic content list (lightning icon). From that list, select the myJSONArray variable:

To create user-friendly tokens for the properties in JSON objects so that you can select those properties as inputs, use the action named Parse JSON before you use the Create CSV table action.

The following screenshot shows the finished example Create CSV table action:

Follow these general steps to add the Data Operations action named Create CSV table .

On the designer, select the Create CSV table action, if not already selected. In the From box, enter the array or expression to use for creating the table.

For this example, select inside the From box, and select the dynamic content list (lightning icon). From that list, select the myJSONArray variable:

To create user-friendly tokens for the properties in JSON objects so that you can select those properties as inputs, use the action named Parse JSON before you use the Create CSV table action.

The following screenshot shows the finished example Create CSV table action:

Customize table format

By default, the Columns property is set to automatically create the table columns based on the array items. To specify custom headers and values, follow these steps:

If the Columns property doesn't appear in the action information box, from the Advanced parameters list, select Columns .

Open the Columns list, and select Custom .

In the Header property, specify the custom header text to use instead.

In the Value property, specify the custom value to use instead.

To return values from the array, you can use the item() function with the Create CSV table action. In a For_each loop, you can use the items() function .

For example, suppose you want table columns that have only the property values and not the property names from an array. To return only these values, follow these steps for working in designer view or in code view.

This example returns the following result:

Apples,1
Oranges,2
Work in designer view
In the Create CSV table action, keep the Header column empty. On each row in the Value column, dereference each array property that you want. Each row under Value returns all the values for the specified array property and becomes a column in your table.
For each array property that you want, in the Value column, select inside the edit box, and then select the function icon, which opens the expression editor. Make sure that the Function list appears selected.
In the expression editor, enter the following expression, but replace <array-property-name> with the array property name for the value that you want. When you're done with each expression, select Add.
Syntax: item()?['<array-property-name>']
Example values:
item()?['Description']
item()?['Product_ID']
Work in code view
In the action's JSON definition, within the columns array, set the header property to an empty string. For each value property, dereference each array property that you want.
On the designer toolbar, select Code view.
In the code editor, find the action's columns array. For each column of array values that you want, add an empty header property and the following expression for the value property:




    

Syntax:
   "header": "",
   "value": "@item()?['<array-property-name>']"
Example:
"Create_CSV_table": {
   "type": "Table",
   "inputs": {
      "columns": [
            "header": "",
            "value": "@item()?['Description']"
            "header": "",
            "value": "@item()?['Product_ID']"
      "format": "CSV",
      "from": "@variables('myJSONArray')"
If you copy the example into your code, make sure you include the runAfter action at the end of the Create_CSV_table definition.
Switch back to designer view to review the results.
Test your workflow
To confirm whether the Create CSV table action creates the expected results, send yourself a notification that includes output from the Create CSV table action.
In your workflow, add an action that can send you the results from the Create CSV table action.
This example continues by using the Office 365 Outlook action named Send an email (V2).
In this action, for each box where you want the results to appear, select inside each box, which opens the dynamic content list. From that list, under the Create CSV table action, select Output.
Save your workflow, and then manually run your workflow. On the designer toolbar, select Run > Run.
If you use the Office 365 Outlook - Send an email action, the following example shows the result:
If your table is returned with incorrect formatting, see how to check your table data formatting.
Create HTML table action
To create an HTML table that has the properties and values from JavaScript Object Notation (JSON) objects in an array, use the Create HTML table action. You can then use the resulting table in actions that follow the Create HTML table action.
To try the Create HTML table action, follow these steps by using the workflow designer. Or, if you prefer working in the code view editor, you can copy the example Create HTML table and Initialize variable action definitions from this guide into your own logic app's underlying workflow definition: Data operation code examples - Create HTML table. For more information about the Create HTML table action in the underlying JSON workflow definition, see the Table action.
Consumption
Standard
Follow these general steps to add the Data Operations action named Create HTML table.
On the designer, select the Create HTML table action, if not already selected. In the From box, enter the array or expression to use for creating the table.
For this example, select inside the From box, and select the dynamic content list (lightning icon). From that list, select the myJSONArray variable:
To create user-friendly tokens for the properties in JSON objects so that you can select
those properties as inputs, use the action named Parse JSON
before you use the Create HTML table action.
The following screenshot shows the finished example Create HTML table action:
Follow these general steps to add the Data Operations action named Create HTML table.
On the designer, select the Create HTML table action, if not already selected. In the From box, enter the array or expression to use for creating the table.
For this example, select inside the From box, and select the dynamic content list (lightning icon). From that list, select the myJSONArray variable:
To create user-friendly tokens for the properties in JSON objects so that you can select
those properties as inputs, use the action named Parse JSON
before you use the Create HTML table action.
The following screenshot shows the finished example Create HTML table action:
Customize table format
By default, the Columns property is set to automatically create the table columns based on the array items. To specify custom headers and values, follow these steps:
If the Columns property doesn't appear in the action information box, from the Advanced parameters list, select Columns.
Open the Columns list, and select Custom.
In the Header property, specify the custom header text to use instead.
In the Value property, specify the custom value to use instead.
To return values from the array, you can use the item() function with the Create HTML table action. In a For_each loop, you can use the items() function.
For example, suppose you want table columns that have only the property values and not the property names from an array. To return only these values, follow these steps for working in designer view or in code view.
This example returns the following result:
Apples,1
Oranges,2
Work in designer view
In the Create HTML table action, keep the Header column empty. On each row in the Value column, dereference each array property that you want. Each row under Value returns all the values for the specified array property and becomes a column in your table.
For each array property that you want, in the Value column, select inside the edit box, and then select the function icon, which opens the expression editor. Make sure that the Function list appears selected.
In the expression editor, enter the following expression, but replace <array-property-name> with the array property name for the value that you want. When you're done with each expression, select Add.
Syntax: item()?['<array-property-name>']
Examples:
item()?['Description']
item()?['Product_ID']
Work in code view
In the action's JSON definition, within the columns array, set the header property to an empty string. For each value property, dereference each array property that you want.
On the designer toolbar, select Code view.
In the code editor, find the action's columns array. For each column of array values that you want, add an empty header property and the following expression for the value property:
Syntax:
   "header": "",
   "value": "@item()?['<array-property-name>']"
Example:
"Create_HTML_table": {
   "type": "Table",
   "inputs": {
      "columns": [
            "header": "",
            "value": "@item()?['Description']"
            "header": "",
            "value": "@item()?['Product_ID']"
      "format": "HTML",
      "from": "@variables('myJSONArray')"
If you copy the example into your code, make sure you include the runAfter action at the end of the Create_HTML_table definition.
Switch back to designer view to review the results.
Test your workflow
To confirm whether the Create HTML table action creates the expected results, send yourself a notification that includes output from the Create HTML table action.
In your workflow, add an action that can send you the results from the Create HTML table action.
This example continues by using the Office 365 Outlook action named Send an email (V2).
In this action, for each box where you want the results to appear, select inside each box, and then select the lightning icon, which opens the dynamic content list. From that list, under the Create HTML table action, select Output.
Save your workflow, and then manually run your workflow. On the designer toolbar, select Run > Run.
If you use the Office 365 Outlook - Send an email action, the following example shows the result:
Filter array action
To create a subset array from an existing array where the subset items meet specific criteria, use the Filter array action. You can then use resulting filtered array in actions that follow the Filter array action.
To try the Filter array action, follow these steps by using the workflow designer. Or, if you prefer working in the code view editor, you can copy the example Filter array and Initialize variable action definitions from this guide into your own logic app's underlying workflow definition: Data operation code examples - Filter array. For more information about the Filter array action in the underlying JSON workflow definition, see the Query action.
Any filter text that you use in your condition is case sensitive. Also, this action
can't change the format or components of items in the array.
For actions to use the array output from the Filter array action, either those
actions must accept arrays as input, or you might have to transform the output array
into another compatible format.
If you call an HTTP endpoint and receive a JSON response, use the
Parse JSON action to process the JSON response.
Otherwise, the Filter array action can read only the response body
and not the structure of the JSON payload.
Although this example uses a simple integer array, this action is especially useful for JSON
object arrays where you can filter based on the objects' properties and values.
Follow these general steps to find the Data Operations action named Filter array.
On the designer, select the Filter array action, if not already selected. In the From box, enter the array or expression to use as the filter.
For this example, select inside the From box, and then select the lightning icon, which opens the dynamic content list. From that list, select the previously created variable:
For the condition, specify the array items to compare, select the comparison operator, and specify the comparison value.
This example uses the item() function to access each item in the array, while the Filter array action searches for array items where the value is greater than one. The following screenshot shows the finished example Filter array action:
Although this example uses a simple integer array, this action is especially useful for JSON
object arrays where you can filter based on the objects' properties and values.
Follow these general steps to find the Data Operations action named Filter array.
On the designer, select the Filter array action, if not already selected. In the From box, enter the array or expression to use as the filter.
For this example, select inside the From box, and then select the lightning icon, which opens the dynamic content list. From that list, select the previously created variable:
For the condition, specify the array items to compare, select the comparison operator, and specify the comparison value.
This example uses the item() function to access each item in the array, while the Filter array action searches for array items where the value is greater than one. The following screenshot shows the finished example Filter array action:
Test your workflow
To confirm whether Filter array action creates the expected results, send yourself a notification that includes output from the Filter array action.
In your workflow, add an action that can send you the results from the Filter array action.
This example continues by using the Office 365 Outlook action named Send an email (V2).
In this action, complete the following steps:
For each box where you want the results to appear, select inside each box, and then select the function icon, which opens the expression editor. Make sure that the Function list appears selected.
To get the array output from the Filter array action, enter the following expression, which uses the body() function with the Filter array action name, and then select Add.
body('Filter_array')
Save your workflow, and then manually run your workflow. On the designer toolbar, select Run > Run.
If you use the Office 365 Outlook - Send an email action, the following example shows the result:
Join action
To create a string that has all the items from an array and separate those items with a specific delimiter character, use the Join action. You can then use the string in actions that follow the Join action.
To try the Join action, follow these steps by using the workflow designer. Or, if you prefer working in the code view editor, you can copy the example Join and Initialize variable action definitions from this guide into your own logic app's underlying workflow definition: Data operation code examples - Join. For more information about the Join action in your underlying workflow definition, see the Join action.
Consumption
Standard
Follow these general steps to find the Data Operations action named Join.
On the designer, select the Join action, if not already selected. In the From box, enter the array that has the items that you want to join as a string.
In the From box, enter the array that has the items you want to join as a string.
For this example, select inside the From box, and then select the lightning icon, which opens the dynamic content list. From that list, select the previously created variable:
Follow these general steps to find the Data Operations action named Join.
On the designer, select the Join action, if not already selected. In the From box, enter the array that has the items that you want to join as a string.
For this example, select inside the From box, and then select the lightning icon, which opens the dynamic content list. From that list, select the previously created variable:
Test your workflow
To confirm whether the Join action creates the expected results, send yourself a notification that includes output from the Join action.
In your workflow, add an action that can send you the results from the Join action.
This example continues by using the Office 365 Outlook action named Send an email (V2).
In this action, for each box where you want the results to appear, select inside each box, and then select the lightning icon, which opens the dynamic content list. From that list, under the Join action, select Output.
Save your workflow, and then manually run your workflow. On the designer toolbar, select Run > Run.
If you use the Office 365 Outlook - Send an email action, the following example shows the result:
Parse JSON action
To reference or access properties in JavaScript Object Notation (JSON) content, you can create user-friendly fields or tokens for those properties by using the Parse JSON action. That way, you can select those properties from the dynamic content list when you specify inputs for subsequent actions in your workflow. For this action, you can either provide a JSON schema or generate a JSON schema from your sample JSON content or payload.
To try the Parse JSON action, follow these steps by using the workflow designer. Or, if you prefer working in the code view editor, you can copy the example Parse JSON and Initialize variable action definitions from this guide into your own logic app's underlying workflow definition: Data operation code examples - Parse JSON.
For more information about this action in your underlying workflow definition, see Parse JSON action.
Consumption
Standard
In the Azure portal or Visual Studio Code, open your logic app workflow in the designer.
This example uses the Azure portal and a sample workflow with a Recurrence trigger followed by an Initialize variable action. This action is set up to create a variable whose initial value is the following JSON object with properties and values:
    "Member": {
        "Email": "Sophia.Owen@fabrikam.com",
        "FirstName": "Sophia",
        "LastName": "Owen"
Follow these general steps to find the Data Operations action named Parse JSON.
On the designer, select the Parse JSON action, if not already selected. In the Content box, enter the JSON object that you want to parse.
For this example, select inside the Content box, and then select the lightning icon, which opens the dynamic content list. From that list, select the previously created variable:
In the Schema box, enter the JSON schema that describes the JSON object, or payload, that you want to parse.
For this example, use the following schema for the sample JSON object:
    "type": "object",
    "properties": {
        "Member": {
            "type": "object",
            "properties": {
                "Email": {
                    "type": "string"
                "FirstName": {
                    "type": "string"
                "LastName": {
                    "type": "string"
If you don't have a schema, you can generate the schema from the JSON object:
In the Parse JSON action, under the Schema box, select Use sample payload to generate schema.
In the Enter or paste a sample JSON payload box, enter the JSON content or object, and then select Done, for example:
    "Member": {
       "Email": "Sophia.Owens@fabrikam.com",
       "FirstName": "Sophia",
       "LastName": "Owens"
In the Azure portal or Visual Studio Code, open your logic app workflow in the designer.
This example uses the Azure portal and a sample workflow with a Recurrence trigger followed by an Initialize variable action. This action is set up to create a variable whose initial value is the following JSON object with properties and values:
    "Member": {
        "Email": "Sophia.Owens@fabrikam.com",
        "FirstName": "Sophia",
        "LastName": "Owens"
Follow these general steps to find the Data Operations action named Parse JSON.
On the designer, select the Parse JSON action, if not already selected. In the Content box, enter the JSON object that you want to parse.
For this example, select inside the Content box, and then select the lightning icon, which opens the dynamic content list. From that list, select the previously created variable:
In the Schema box, enter the JSON schema that describes the JSON object, or payload. that you want to parse.
For this example, use the following schema for the sample JSON object:
    "type": "object",
    "properties": {
        "Member": {
            "type": "object",
            "properties": {
                "Email": {
                    "type": "string"
                "FirstName": {
                    "type": "string"
                "LastName": {
                    "type": "string"
If you don't have a schema, you can generate the schema from the JSON object:
In the Parse JSON action, under the Schema box, select Use sample payload to generate schema.
In the Enter or paste a sample JSON payload box, enter the JSON content or object, and then select Done, for example:
    "Member": {
       "Email": "Sophia.Owens@fabrikam.com",
       "FirstName": "Sophia",
       "LastName": "Owens"
Test your workflow
To confirm whether the Parse JSON action creates the expected results, send yourself a notification that includes output from the Parse JSON action.
In your workflow, add an action that can send you the results from the Parse JSON action.
This example continues by using the Office 365 Outlook action named Send an email (V2).
In this action, for each box where you want the results to appear, select inside each edit box, and then select the lightning icon, which opens the dynamic content list. From that list, under the Parse JSON action, select the properties from the parsed JSON object.
This example selects the following properties: Body FirstName, Body LastName, and Body Email
Save your workflow, and then manually run your workflow. On the designer toolbar, select Run > Run.
If you use the Office 365 Outlook - Send an email action, the following example shows the result:
Select action
By default, the Select action creates an array that contains JSON objects built from the values in an existing array. For example, you can create a JSON object for each value in an integer array. Specify the properties that each JSON object must have and map the values from the source array to the properties. Although you can change the component JSON objects, the output array always has the same number of items as the source array. To use the output array from the Select action, subsequent actions must either accept arrays as input, or you might have to transform the output array into another compatible format.
To try the Select action, follow these steps by using the workflow designer. Or, if you prefer working in the code view editor, you can copy the example Select and Initialize variable action definitions from this guide into your own logic app's underlying workflow definition: Data operation code examples - Select. For more information about this action in your underlying workflow definition, see Select action.
For an example that creates an array with strings or integers built from the values in a JSON object array,
see the Select and Initialize variable action definitions in
Data operation code examples - Select.
Follow these general steps to find the Data Operations action named Select.
On the designer, select the Select action, if not already selected. In the From box, enter the source array that you want to use.
For this example, select inside the From box, and then select the lightning icon, which opens the dynamic content list. From that list, select the previously created variable:
For the Map property, in the left column, enter a property name to describe all the values in the source array.
This example uses Product_ID as the property name for the integer values in the source array. So, each value, for example, 1, 2, 3, and 4, is a product ID.
In the right column, enter an expression that specifies the value in the source array for the property name in the left column.
This example uses the item() function to iterate through and access each item in the array.
Select inside the right column, and then select the function icon, which opens the expression editor. Make sure that the Function list appears selected.
In the expression editor, enter the function named item(), and then select Add.
Follow these general steps to find the Data Operations action named Select.
On the designer, select the Select action, if not already selected. In the From box, enter the source array that you want to use.
For this example, select inside the From box, and then select the lightning icon, which opens the dynamic content list. From that list, select the previously created variable:
For the Map property, in the left column, enter a property name to describe all the values in the source array.
This example uses Product_ID as the property name for the integer values in the source array. So, each value, for example, 1, 2, 3, and 4, is a product ID.
In the right column, enter an expression that specifies the value in the source array for the property name in the left column.
This example uses the item() function to iterate through and access each item in the array.
Select inside the right column, and then select the function icon, which opens the expression editor. Make sure that the Function list appears selected.
In the expression editor, enter the function named item(), and then select Add.
Test your workflow
To confirm whether the Select action creates the expected results, send yourself a notification that includes output from the Select action.
In your workflow, add an action that can send you the results from the Select action.
In this action, complete the following steps:
For each box where you want the results to appear, select inside each box, and then select the function icon, which opens the expression editor. Make sure that the Function list appears selected.
To get the array output from the Select action, enter the following expression, which uses the body() function with the Select action name, and select Add:
body('Select')
Save your workflow, and then manually run your workflow. On the designer toolbar, select Run > Run.
If you use the Office 365 Outlook - Send an email action, the following example shows the result:
Troubleshooting
Format table data
If your CSV table or HTML table is returned with incorrect formatting, make sure your input data has line breaks between rows.
Incorrect formatting:
Fruit,Number Apples,1 Oranges,2
Correct formatting:
Fruit,Number
Apples,1
Oranges,2
To add line breaks between rows, add one of the following expressions to your table:
replace(body('Create_CSV_table'),'','<br/>')
replace(body('Create_HTML_table'),'','<br/>')
For example:
	"Send_an_email_": {
		"inputs": {
			"body": {
				"Body": "<p>Results from Create CSV table action:<br/>\n<br/>\n<br/>\n@{replace(body('Create_CSV_table'),'\r\n','<br/>')}</p>",
				"Subject": "Create CSV table results",
				"To": "sophia.owen@fabrikam.com"
Related content
Managed connectors for Azure Logic Apps
Built-in connectors for Azure Logic Apps