The wf given below correctly converts table data to a json output, until n8n v1.42.1 (as seen here):
But v1.43.1+ onwards, the conversion functionality seems broken:
Hi @shrey-42 ,
Thanks for pointing this out! 
I’ve retraced you steps and got the same error. And we did do some changes to the Extract from File Node on version 1.43.0.
I’ll raise this internally and we’ll take a look at it!
Hey @shrey-42,
Just looking at this… You are using 2 different nodes, Does the old node still work as expected on the newer version?
Do you have the actual worklfow available that you are using, I would expect the HTML node to be used to convert HTML text to items like you are doing so something went wrong at some point.
Hey @Jon
The first screenshot uploaded here wasn’t correct. Have updated it now.
Nope (as seen in the 2nd screenshot).
[btw, if by the old node you mean the original Spreadsheet node, that isn’t working v1.43.1 onwards either]
The test wf shared above is pretty much what i’m using actually.
Hey @shrey-42,
Perfect thanks, So in your full worklfow are you using the code node as well?
The last change was for the CSV reading not the HTML so I don’t think the change to the node is related but I did notice the output shows new line characters so I wonder if a core change around those is causing the issue.
Yes. The html data has been copied from a source and being used as the origin using a Code node.
In the HTML you were using was it just the table or was it the full markup?
Just had a proper play looks like it is something between 1.43 and 1.43.1 so the change is somewhere here: Comparing [email protected]@1.43.1 · n8n-io/n8n · GitHub
I suspect it could be the sheetjs upgrade one, I would recommend using the HTML node to extract the fields you need which would be a better solution and we can update the release notes to mention a breaking change.
just the table.
The output in that case is not comparable (nor relevant) to the original json output at all.
Although, I seem to have found the source of this issue:
The ‘HTML file’ generated by the Convert to File node (SheetJS Export) is broken (HTML characters have been converted to HTML entities), which is subsequently not properly interpreted by the Extract from File node, thus yielding improper output.
Additionally, the Convert to File node would be expected to have a field to denote the incoming json property containing the data to be written to the file.
Here’s the modified example (with the full html).
This is the html content of this generated file →
Whereas, this is the expected output (and which is yielding proper results with the Extract from File node) →
Is there any fix for this in the pipeline?
Hello @shrey-42,
‘Convert to HTML’, and other operations that work with spreadsheets, convert incoming inputs data into a table where each item became row and them from this data binary file is created
As you want to use string data you need to use ‘Convert to Text File’ operation
I did add workflow with two branches- first with fix and other of how to use ‘Convert to HTML’ operation to convert incoming data into html file
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
