Attachment content cannot be null or empty

[abhender]

Hi all,

I know this question has been asked before but there was no solution so I figured I'd ask again.

I built a simple but fairly long flow that takes the information from a SharePoint list entered from Powerapps, translates it, then creates two different pdf's (Spanish and English) and sends it out in an email to multiple recipients. 

The flow works 50% of the time with the other 50% occurring at the very last step [send an email(V2)], giving me the error:

Parameter 'Attachment Content' cannot be null or empty.
clientRequestId: 37adc4f7-288a-4616-b4f8-cdd7cf65835b

The attachment content is not null, I have dynamic files running, there is nothing hard coded, I have even tried adding delays and switching around the flow. 

The only temporary fix is building the flow from scratch each time which is terribly time consuming. 

Any help is appreciated! TIA

[jinivthakkar]

Can you please share screenshots ? and if there are multiple attachments are you iterating in a loop and then getting attachment content for each file ?

Please check this link : 

https://www.techmeet360.com/blog/how-to-send-an-email-attachment-using-microsoft-flow/

[HEATFreight]

I have an attachment which is clearly not empty, yet my flow has an email error.

"body":{"status":400,"message":"Parameter 'Attachment Content' cannot be null or empty.\r\nclientRequestId: 2c948bd4-b51d-4097-861f-44dfdccb8d14","error":{"message":"Parameter 'Attachment Content' cannot be null or empty."},"source":"office365-wus.azconn-wus-01.p.azurewebsites.net"}}

The output from the "Get file content using path" action is this

{ "$content-type": "application/pdf", "$content": "JVBERi0xLjQKJaqr....=" }

And the attachment is attached using the normal way (not as an array). Like

Name.pdf

and

@body('Get_file_content_using_path')?['body']

How can this be null? Is this a transient issue that will fix itself? I worry that I am over my OneDrive data quota, and not sure how to fix.

[HEATFreight]

Ok now I have passed in the attachment as an array via a compose function:

{
  "Name": @{outputs('Get_file_metadata_using_path')?['body/DisplayName']},
  "Content": @{body('Get_file_content_using_path')?['body']}
}

And the compose:

@{variables('Array')}

Attachments (Send an email V2):

[
  {
    "Name": "Name.pdf",
    "Content": {
      "$content-type": "application/pdf",
      "$content": "JVBERi0xLjQKJaqrrK0KMSAwIG9iago8PAo...="
    }
  }
]

 
Error details:

The 'inputs.parameters' of workflow operation 'Send_an_email_(V2)' of type 'OpenApiConnection' is not valid. Error details: The API operation 'SendEmailV2' is missing required property 'emailMessage/Attachments/0/ContentBytes'.

[gpohlod]

Hi @abhender ,

Are you sending one email or multiple emails at the same time?

If multiple emails, try adding a delay of 20 seconds between emails and see if that helps. It seems to have solved my issue. 

[HEATFreight]

Hey @abhender, I solved my problem by figuring out the difference between "Content" and "ContentBytes".

Certain scenarios require one or the other, and I don't really understand it, but I know when to use each one for my use cases. If you are having trouble, switch to the other and see if it fixes the problem, like:

{
  "Name": @{outputs('Get_file_metadata_using_path')?['body/DisplayName']},
  "Content": @{body('Get_file_content_using_path')?['body']}
}

vs.

{
  "Name": @{outputs('Get_file_metadata_using_path')?['body/DisplayName']},
  "ContentBytes": @{body('Get_file_content_using_path')?['body']}
}

[goodingip]

I am getting the same issues, sending one email with 2 attachments. Both attachments are valid, as saving them to OneDrive works consistently immediately before the send as email. I tried the suggestion from @HEATFreight; the default in my PowerApps is to use "ContentBytes": for the attachment, and it gives me an error if I change it to "Content":  I am wondering if it could be a timeout issue, as it seems more likely when the attachments are larger (around 2Mb in my task - a PDF). It also works fine sometimes, and not others, with no change to the script. 

[gpohlod]

Hey @goodingip ,

I found it's very inconsistent and the only way I've been able to get it working consistently was to build it as a solution that calls child flows. The parent flow passes the path info of the file to the child flow which gets it from in my case SharePoint, then attaches the email and sends and returns to the parent flow, so it can call the next email. 

[abhender · Accepted solution]

Thanks everyone for the replies, the flow started working shortly after I posted this message. The fix was that I had to upload all of the translated versions (documents) to the SharePoint list first as an attachment. Unlike before, I just had my array pulling the generated files as soon as they became translated. Then I used the whole "get attachments" scenario to grab them all again and put them in an array to be emailed. (There are some skipped steps below but they're not pertinent to the flow working). Hope this helps.

[goodingip]

That's helpful advice, thanks @gpohlod. I think what is happening in our case is that the licence I'm using isn't meaty enough (PowerApps Plan 2), so we're getting arbitrarily cut off when we do something complex. I've see it randomly stop at other points in our flow, so in our case it doesn't seem to be anything per se to do with the email attachments. 

See Limits and configuration - Power Automate | Microsoft Docs for how Microsoft have chosen to limit flow functionality unless you buy a per-flow plan. I borrowed a Dynamics 365 licence and it instantly fixed the problem. As a charity, I'm trying to do stuff without fancy licencing, but in this case we're going to have to pay to get the business efficiency we are looking for. 

[everstab]

Hi,

I am getting the empty content error message when trying to attach attachments from a SharePoint list to an approval, even though the attachments are NOT empty.  I believe my flow is correct and it does work up until the approval.  I even put in a 5 minute delay since I work in a non-profit and some posts said our version of Power Automate may not be as robust in performance.  I am still getting the same error.  Any help would be greatly appreciated!   

Thanks,

Pat

[HEATFreight]

@everstab Pat, I don't have time to inspect your flow logic, but when I have problems with this I simply change "Content" to "ContentBytes" in the 'Append to array variable' action.

Certain types of files require one format and other kinds of files require the other format. If neither of these work for the files you are working with, then I would need to spend more time researching your specific use case.

PDF files are encoded as a stream of bits. Zeros and ones. So you need the "ContentBytes" tag instead of "Content" in your array. Photos on the other hand require that same "Content" tag you're already using. You need a conditional statement to branch the logic between different filetypes, and append the file content to the array using the correct content tag for that type of file.

[zeedee11]

I was able to solve this issue by using the sharepoint action "get file content", then putting file content into a compose action, and then putting the outputs of the compose action as the attachment content. Seems like a silly workaround but it worked for me.

[HEATFreight]

@zeedee11 yeah, I use that trick quite a bit actually... When in doubt, compose it out!

Compose is also great because when you look at the data for a flow run it lets you see the current value of the entity which you're composing at that particular point in the flow, even when you would otherwise not be able to. It lets you pull some data out in a way that's easily visible but without having to initialize and set a variable. Love the compose action! Just wish I knew how to use the Parse JSON action...

Rant:
I have some unstructured JSON from a JotForm flow trigger that I need to dynamically parse for key-value pairs. The JotForm connector for Power Automate is poorly built such that the dynamic content keys for all fields in the JotForm are appended with a number which is just their order in the form, and this dynamic content breaks if that number changes. So if you rearrange the order of the fields in the JotForm, or add or remove fields, that changes the number appended to the field name which becomes the key name for that dynamic content, and when the number changes the reference link to that dynamic content breaks. So, for instance, if you were to delete the first field on a JotForm, the entire flow and every single variable coming from JotForm would break because the numbers appended to the dynamic content would all be off by one and thus none of the variables would populate with the data. So if —instead of initializing each variable with the dynamic content supplied by the JotForm connector— we were to parse the unstructured JSON in the JotForm flow trigger response body to populate variables with the correct value regardless of the number prepended to the key name, that would completely eliminate the issue. We could change up the order of the JotForm fields any way we like and the flow would always be able to match the correct value to its corresponding flow variable. The only thing that would break it would be removing a field entirely or changing the name of a field in the JotForm. Otherwise the flow would be robust in the face of changes to the JotForm. But until I can figure out how to use the Parse JSON action or some other Power Automate trick, I am stuck with an extremely brittle connection to JotForm which is not resilient to changes in the order of fields on the form. Ok rant over.

Anyone knows how to parse unstructured JSON in Power Automate, I'm all ears!

EDIT: Nvm, but holy smokes I figured out a solution to fix JotForm's un-parsable JSON. I filtered the "## - " section that prepends the key and i figured out how to recursively addproperty() to a JSON object variable until my JotForm triggerbody was rebuilt without the asinine numbers prepending my keys, and that can now be quite robustly parsed with Parse JSON. All I gotta do is  keep the JotForm field names the same, and add new schema every time I have a new form field so that I can dynamically grab it as a Parse JSON dynamic content. I can tell you that this sort of magic basically demands extensive use of Compose. Compose is the MVP of Power Automate!