I have a workflow that successfully creates a PDF using the Report PDF Writer.
However, I am noticing it can only save the the local machine. there is no way to use the SMB Connector or SharePoint connector to save the resulting file to anywhere else. I don’t even see a way to change the relative location of the file so this works on the Data Hub.
Are there other nodes that can generate a PDF file that allow you to save to another location?
@aehrenworth can you write to a local drive and then upload (transfer) the file?
Hi @aehrenworth ,
full Filehandling support in the PDF Writer is on our todo list (internal reference UIEXT-1705).
In its current state, you can use “KNIME URLs” to make it write to the Hub file system:
knime://knime.workflow/../report.knime to write the report next to the workflow wherever it is located: on Hub if it’s located on Hub, in your local space if it’s located in your local Analytics Platform – regardless of whether it’s being executed on Hub as a job or on your local AP as a temporary copy (cloud icon in tab)Here’s a small example demonstrating the use: Forum-87117-PDF Writer KNIME URL – KNIME Community Hub
Once it’s written next to the workflow or in some other temporary place on Hub, you should be able to move/copy it with the Transfer Files/Folder node.
Maybe I am doing something wrong, but when I look at each example in the workflow, they all show my personal folder with report.pdf-- do I need to make the adjustments myself based on the “comments” under each node?
No, adjustments should not be needed. 
The “Output location” should contain the knime:// URLs as mentioned in the comments.
What version of AP are you using? I found a fixed regression ticket describing this behavior, but that was only observed in the 5.4 RC, not for the 5.3 line and not for the actual 5.4 release.
(so the regression was never in a released product as far as we are aware)
It also only resets the values if you open the dialog. If you import the workflow and never open the dialog, it will still contain the knime:// URLs.
Ok, I will try reloading without opening the dialog. We are on 5.3.2 and there were multiple errors/prompts when I first opened it.
Maybe that explains it. I created it with 5.4. It is possible that 5.3 does not understand some of the new things added in 5.4 and resets to some default values.
I’ll create a 5.3 version.
This should work in 5.3.
AP 5.3 definitely does not understand the location field created by 5.4 (it’s not “forward compatible”).
Excellent thanks!
One other question I noticed the files that are getting created by the node are sometimes VERY large based on the content. If an image is included could that be the main reason for the large file size of of the PDF output?
I’m not familiar with PDF internals, but I think to remember images are embedded in the PDF as-is (though I think some PDF creator tools allow to set image compression). So it would not be surprising that a large number of images or large images increase filesize. Are you including many or just a few large images? Do you create the images in another software or using Views?
If you use the “new” Views, you can set them to emit SVGs (since 5.4), which should dramatically reduce file size. Unfortunately, the Generic ECharts View does not support that yet, though.
I am just doing one image for now per PDF output but they may be of higher resolution / quality because it is being passed through an AI image process first (and needs to be as detailed as possible). let me see if I can reduce the image size before passing it over. we are on 5.3.2 and may not upgrade until the second half of 2025 so need an interim solution.
If the images are not generated by the Views themselves (i.e. are from this AI process), the upgrade will not help. The SVG output is only for Views such as Bar Chart, etc. but not Image View.
Instead of reducing the image dimensions, you could also try to optimize/compress them.
I have personally used pngcrush in the past to reduce the file size of PNGs, for JPEGs there seems to be jpegoptim.
Maybe the KNIME Image extension also offers image compression functionality, but I have no experience with it yet.
Another route would be a Python script node calling an appropriate library. pillow or OpenCV come to mind. And then there is also imagemagick which could help 
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.