PRes Connect Workflow Slow

We run some jobs that process a CSV file using a PReS Connect workflow. The workflow loops over each line in the CSV and creates a pdf using a PP Design document (7.6.2) then outputs using an LPR printer queue output. The main issue we are having is that the PDF creation takes between 7-10 seconds…for a 1 to 2 page PDF file. This is a big issue with a 2000 row CSV file taking several hours to generate the documents.

The design document contains a few background images(pdfs of between 70kb and 200kb in size, quite small) and depending on the data will create a 1 or 2 page PDF made up of the appropriate pages from the design. There are 5 different pages in the design that could be included in this PDF.

I tried outputting the PDF to file system instead of printer, this is just as slow.
As a test I tried a test data set with around 500 records and outputting the entire document as one PDF file, this takes sub 30 seconds to generate a 1000 page PDF. Unfortunately we cant run this way.

Unsure where to start with improving this processs, if there is a better way, what I am doing wrong…any guidance would be much appreciated.

Why don’t you generate the whole thing into 1 PDF and then split said PDF into all single files?

There is actually another small part in the workflow a branch, which sends a pre generated transcript to the printer. So we loop sending the Transcript to the printer(almost instant), then the generate and send the PDF for that data row, repeat till end of file. We need them in this order for pack and dispatch.

Each PDF is between 1 and 2 pages depending on the data so would have to loop and split depending on that.

I will see if I can incorporate splitting into the workflow, unsure how to do it…

When you say …pre-generated transcript…

What is it?
A fixed one?
Based on the splitted CSV data?

The “transcript” is just a single page PDF file that is associated with the record in the CSV and its already available to the process so we just load the file and send to print queue.

I have been playing with generating the whole PDF and splitting it. Turns out that it is not variable length PDF files, there are always 2 pages so I can generate a whole document, split and send to a folder quickly and easily.

The issue now is queueing the transcript then the split pdf file, I am experimenting with some kind of meta data in the PDF so I can tell what to name the file when splitting to folder and then I loop through the CSV again queueing the transcript then the named PDF. Fun fun