Power Automate Flow - Convert File to PDF not working

[HaveeLand]

July 10, 2024

Hello All,

Short Version of issue: 

• Flow runs successfully.
• Flow creates an .html file
• Flow converts .html file to .pdf
• Flow creates a .pdf file using the "body" of the conversion from .html to .pdf
• 2 files are created in my OneDrive account.
  • .html file opens perfectly (1st File)
  • .pdf file Does Not open (2nd File)

Longer version if you need more details.

It appears that a few years ago, other people were having a similar issue across multiple users and the problem eventually was corrected after a few days.  Hopefully the issue is only on my account and not a global one like last time.

Here is my issue:  I have a flow that is Manually triggered and I write some basic "Hello World" HTML code in a "Compose" function. 

I use {Create file} and I am able to create an .html file in my OneDrive folder and the document opens in any browser I use. 

HOWEVER, when I use the {Convert file} for a PDF target type, using the html file from above, the flow runs successfully, however in the raw output file, it shows a "Content-Length": "0" (zero) value. 

Which then is considered the {body} and when I use it in the {Create file} under the File Content to convert the file to PDF. It generates a PDF file with a value of 1kb and I get an error message that the PDF file can NOT be opened.

So, my TEST Flow creates 2 files.  An .html file that works as intended and it also creates a .pdf file that I am NOT able to open.

And yes... as of this afternoon 7/10/24, everything was working and converting properly.

So, hopefully I added all of the information needed to solve the issue and you can direct me to a solution.

Thank you all in advance

-Javi

[Sayan]

Hey @HaveeLand 

in your workflow, you are trying to do: 

• Creating the HTML file.
• Converting the file to PDF, which means the HTML code is converted to PDF. That's why the .pdf format does not support HTML content.

Solution: 

After creating the HTML file, get the content of the file and convert it to PDF, which runs the HTML code and converts it to PDF. After that, use the output to create the file, and then you will be able to open the PDF file.

Workflow:

 Output:

Hope it will also work for you. 😊

------------------------------------------------------------------------------------------------------------------------------

If my suggestion helped you, please give it a Thumbs up 👍 and Mark it as a Solution ✔️ .

This motivates to solve more question.

Thanks,

Sayan Patra

RPA Developer

[bestrya]

I am having a similar issue with Onedrive for Business - Convert to PDF.  It had been working fine until the afternoon of 7/10/2024. The flow is supposed to convert an html file to pdf, but the pdf is now always 1kb in size and cannot be opened (Adobe Acrobat error attached). I did some testing with the connector and was able to successfully convert the html to a jpg... but will no longer convert to pdf.

[Sayan]

Hey @bestrya 

Follow the above solution. Hope it will work.

Thanks,

Sayan Patra

RPA Developer

[bestrya]

I'm sorry, I initially misread your post.

I went back and added 'Get file content using path', but am now getting an error on 'Convert file'.  If you have any insight into fixing the error, I would appreciate it!

[bestrya]

I just noticed that 'Convert file' is asking for the unique identifier of the file, but I was using 'File content'.  I will try to find the unique identifier for the file and use that.

EDIT: I'm not having much luck. Can you tell me what you used as the "File" input in your "Convert HTML File to PDF" step?

 

[mpfeng]

@bestrya  For the File parameter, use the "Id (The unique identifier of the file or folder)", which is an output of the previous step: the OneDrive Create (html) file. This was working until yesterday afternoon, however, since then, it creates a 4 bytes PDF file with no content, as @HaveeLand has reported.

[HaveeLand]

Hello All,

Thank you for the feedback.  

I was concerned that it was going to be a system issue. Based on @bestrya and @mpfeng it seems like it is.  

However, I like @Sayan suggestion. I have not needed to do that in the past as my flows were working but it makes sense.  Especially if MS changed something. I will try it when I get to the office.  

I will keep you all updated.  

Thanks again 

@HaveeLand 

[Sayan]

Hey @HaveeLand if my suggestion helped you, please give it a Thumbs up 👍 and Mark it as a Solution ✔️ .

This motivates to solve more question.

Thanks,

Sayan Patra

RPA Developer

[ademi]

Hi there, noticed the same problem this morning. The work around does not seem to work because cant figure out where the .html is hiding to use as "path" I tried root but that didnt work, neither did "id". Any chance this is a MS thing that will be sorted out by tomorrow morning?

[HaveeLand]

Well @ademi Unfortunately, last time this happened it took a while for MS to recognize it and then to fix it.  It will all depend on what actually caused it.

[mpfeng]

I tried to follow the steps suggested by @Sayan , but I still get the PDF file with size = 4 bytes with no content.

Notice that the "File" parameter of the "Convert file (Preview)" action expects an identifier of the file, so the "File content" (the output of the "Get file content using path") is not admitted as a parameter.

[WildOrange]

I'm having the exact same issue and that's how I found this post. Wasted hours trying to figure out why I can no longer output PDFs.

Guess I need to rethink using convert to PDF if it could fail on me during critical moments.

[HaveeLand]

@WildOrange I feel your pain.  Until there is a work around from MS, we are stuck.  So here are a couple of options you might have...

1. You can definitely use a different connector or action like "Encodian" or other 3rd party services.  However, for my needs, I need more than 50 per month.  So, unless I purchase a different plan, those will not work for me. But if you only need a handful of conversions, you may be able to migrate to a 3rd party connector.

2. If your flow creates an html page then converts it to PDF (or attempts as that is the current issue)... you may be able to keep the document as an .html document for the time being.  If you really need, you might then be able to manually convert the webpage using  the save as option on the document or use a different system to automate the process of converting the .html document to pdf.  Originally, I was attaching the WEB PAGE to my emails.  My thought was that the customer (recipient) would be able to view the document, data, invoice, whatever it was in PDF, in a browser which everyone has accessible. 

3. Don't convert the document to PDF at all and simply add it to an email.  That is my current solution.  I was able to incorporate all of the .html code and answers from the Microsoft form into the body of the email.  (Essentially, the body of the email and the body of the PDF were identical - I recreated the PDF in the body of the email) So... What I chose to do was simply keep the information in the body of the email and not create the PDF attachment at this time until the system is working again.  

As such, the email contains the important information, and they can print the email at the current time without the attachment.

So I hope this helps out.  In the meantime, I just keep testing my "Test" flow until it works.

Take care

-Javi

[HaveeLand]

My apologies @bestrya I just noticed you asked me a question.  This is the dynamic content that I use.  Under normal circumstances it would work correctly.  It is the OneDrive connector with the "ID" which is the Unique Identifier.

Hope this is helpful.

[HaveeLand · Accepted solution]

Hello All,

It appears that the issue with the OneDrive connector and converting an .html file to a .pdf file has been resolved.  The same code that was not working previously, is now generating a correct .pdf file.

Here is a snapshot of the PDF file viewed through Adobe (which we could not do before)

So thank you to all of you that took an interest.  If I find out how the problem was resolved (other than just waiting until it worked) I will gladly share it with you all.

For all of you that end up reading this post in the future and are wondering... The Issue was discovered on 7/10/2024 around noon and I verified that the system was working on 7/12/2024 at 1am.  So the problem resolved itself in approximately 31 hours.

Take care and we will talk at the next issue.

[Sayan]

Hey @mpfeng the issue occurs due to the incorrect file path that you have provided in "Convert file" please check once that "Id" what value holds and while creating the file provide the "File Content".

Thanks,

Sayan Patra

RPA Developer

[Sayan]

Yes @HaveeLand  sometimes due to the connection failure we face the issue so in that situation we should create a new connection and use it

Thanks,

Sayan Patra

RPA Developer

[WildOrange]

I can confirm it's working again. It was not user error but a system wide issue.

[mpfeng]

Hi @Sayan The "Convert file (Preview)" has two parameters: File "Id" and Target Type (PDF). A file path is not accepted as parameter. I have not provided a file path to the "Convert file (Preview)", as you can see in the screenshot I provided.