Unanswered

Issues with Parse JSON Schema - Column Names not Showing up in Dynamic Content

(0) Share

Report

Posted on by AR-15041655-0

I'm working on building a Power Automate flow that converts a CSV file into a SharePoint list. I'm using the Parse JSON action and initially generated the schema by copying the output from my Select action.

When I leave the schema as is, the flow gives me an error (see screenshot below). However, the column names appear in the dynamic content—which is exactly what I need.

But when I modify the schema to allow for null values, the flow runs successfully, but the column names no longer show up in the dynamic content.

Has anyone run into this before? Is there a way to keep the column names visible in dynamic content while still handling null values properly? I'm not sure what to adjust next.

Here is the error with the schema that does not adjust for null values but I still get the dynamic content :

When I adjust for nulls, the flow works but there is no dynamic content:

Categories:

Building flows

Screenshot 2025-0...

I have the same question (0)

All responses (9)

Answers (1)

Pstork1 68,697 Most Valuable Professional on at

Like (0)

Report

When you adjust to allow the null data type do you also remove those columns from the required list at the end of the schema? I've adjusted the schema before to allow for null values and never had an issue with using the dynamic content later. But you do need to remove them from the required list.

----------------------------------------------------------------------------------
If this Post helped you, please click "Does this answer your question" and give it a like to help others in the community find the answer too!

Paul Papanek Stork, MVP
Blog: https://www.dontpapanic.com/blog

Was this reply helpful? Yes No
Suggested answer

David_MA 12,966 Super User 2025 Season 2 on at

Like (0)

Report
When you adjust the scheme for nulls, I believe you need to format it like this:

{ "type": "object", "properties": { "odata.metadata": { "type": [ "string", "null" ] }, "value": { "type": "array", "items": { "type": "object", "properties": { "odata.type": { "type": [ "string", "null" ] }, "odata.id": { "type": [ "string", "null" ] }, "odata.editLink": { "type": [ "string", "null" ] }, "Title": { "type": [ "string", "null" ] }, "Prio": { "type": [ "integer", "null" ] }, "IPAddress": { "type": [ "string", "null" ] } }, "required": [ "odata.type", "odata.id", "odata.editLink", "Title", "Prio", "IPAddress" ] } } } }

Note how you have the type values within { } rather than [ ] as in my code. This format works for me.

Was this reply helpful? Yes No
AR-15041655-0 12 on at

Like (0)

Report

@Pstork1 I did that, and I am still running into the same issue:

Here is the whole thing:

{

"type": "array",

"items": {

"type": "object",

"properties": {

"First Name": {

"type": [

"string",

"null"

]

},

"Last Name": {

"type": [

"string",

"null"

]

},

"Email": {

"type": [

"string",

"null"

]

},

"External Title": {

"type": [

"string",

"null"

]

},

"Business Unit Name": {

"type": [

"string",

"null"

]

},

"Market Name": {

"type": [

"string",

"null"

]

},

"Location Name": {

"type": [

"string",

"null"

]

},

"Language\r": {

"type": [

"string",

"null"

]

}

}

}

}

Was this reply helpful? Yes No
Pstork1 68,697 Most Valuable Professional on at

Like (0)

Report

What type of action are you trying to input the dynamic content into? The best test would be a Compose because that doesn't do any filtering on data types. It may be that its filtering out the content because its not guaranteed to be a string data type.

----------------------------------------------------------------------------------
If this Post helped you, please click "Does this answer your question" and give it a like to help others in the community find the answer too!

Paul Papanek Stork, MVP
Blog: https://www.dontpapanic.com/blog

Was this reply helpful? Yes No
David_MA 12,966 Super User 2025 Season 2 on at

Like (0)

Report

I think this may be the issue:

  "Language\r": {

  "type": [

  "string",

  "null"

The \r is not valid.

Was this reply helpful? Yes No
Chriddle 8,436 Super User 2025 Season 2 on at

Like (0)

Report
Dynamic Content does not work with multiple types.

Instead of manipulating the schema (as described by Pstork1), you can also ensure that no null value occurs during list creation:

coalsece(<your_value>, '')

Was this reply helpful? Yes No
AR-15041655-0 12 on at

Like (0)

Report

@Chriddle Do I do this in my Select action?

If so, how would i format it? this is how my map fields are formatted:

Left:

outputs('Compose_-_Column_Names')[1]

Right:

split(item(),',')?[1]

Was this reply helpful? Yes No
Pstork1 68,697 Most Valuable Professional on at

Like (1)

Report

If I understand correctly what you are trying to do, then you shouldn't need to handle Nulls because any missing items from the CSV would be blanks, not nulls. I tried reproducing your setup and here's what I got.

1) Used a manual trigger and put the CSV data lines in a variable and a linefeed variable

2) Used Split with the Linefeed variable to create an array of lines from the CSV

3) Used a Data Select to Split each line into fields and load them with the column names Each field is retrieved with an index using

split(item(),',')[0] through split(item(),',')[3]

4) Did a regular Parse JSON with no changes to the Schema

5) All fields now available as dynamic content

----------------------------------------------------------------------------------
If this Post helped you, please click "Does this answer your question" and give it a like to help others in the community find the answer too!

Paul Papanek Stork, MVP
Blog: https://www.dontpapanic.com/blog

Was this reply helpful? Yes No
Verified answer

Chriddle 8,436 Super User 2025 Season 2 on at

Like (0)

Report
In fact, a valid CSV file cannot contain any null values.

Depending on how you convert the CSV it might be because of a blank line at the end of the file.
Simply remove empty lines with a filter before CSV conversion:

Filter array

From:

split( outputs('Compose-CSV'), decodeUriComponent('%0A') )

Filter:

trim(item())

is not equal to

string(null)

Select

From:

skip( body('Filter_array'), 1 )

Name:

split(item(), ',')[0]

Email:

split(item(), ',')[1]

Dob:

split(item(), ',')[2]

Pro tip ;)

Don't use the question mark (optional value) here: split(item(), ',')?[2]

Rather than proceeding with invalid data, you probably want your flow to fail.

Additional advantage: The Select already displays a descriptive error message.

Was this reply helpful? Yes No