Midrange News for the IBM i Community

PDF for You and i Published by: Bob Cozzi on 15 Feb 2011 view comments(6)
© 2011 Robert Cozzi, Jr. All rights reserved.

© Robert Cozzi, Jr. All rights reserved. Reproduction/Redistribution prohibited.
A midrangeNews.com Publication

PDF for You and i

Print Directly to a PDF File on the IFS

NOTE: RPG Report Free Preview Has Ended

Our two February issues require no membership, our two March issues require a free membership, and all subsequent issues of RPG Report require a Premium (paid) membership. Learn more about our membership choices here.

Around V4R2 a small group of former IBM Rochester Engineers retired and started a small software company. That company's purpose was to bring non-traditional interfaces to the IBM i platform. They successfully created or ported everything from a graphical user interface, 5250 emulation, zip file compression and time synchronization. One of their biggest sellers was a driver that allows you to move a SPOOL file to a specific output queue; and output queue that magically transformed the SPOOL file into a PDF file. It works like magic and is still available from the company's website.

Native PDF Print Support

Jumping ahead to today; when IBM i OS v6r1 was shipped, IBM provided its own built-in PDF driver. Unlike the output queue approach described above (which continues to be viable today and there is still no IBM PDF alternative) this new IBM PDF driver works in conjunction with Print File attributes. Just like the PC, to create a PDF file from a document, you override the printer driver and write to the output device. This is done on IBM i by issuing an OVRPRTF command or by creating a print file that includes the PDF attributes. To me, the OVRPRTF command is the obvious choice for this functionality.

Sponsored by: BCD

PDF Prerequisites

To utilize the IBM i PDF driver, you need to make sure you've installed the TS1 licensed program. If you're running v6r1 the product ID is 5661-TS1, if you're running v7r1 or later the product ID is 5770-TS1. You will need to install both the *BASE support and option 1. Here's the summary:

v6R1m0 - 5661-TS1

v7R1m0 - 5770-TS1

*BASE IBM Transform Services for i5/OS
 1    Transforms - AFP to PDF Transform

If you have these prereqs installed, you are ready to start generating PDFs on your system today. In addition to the above software packages, authority to issue the OVRPRTF command and access an IFS folder is required.

Similar to directing output to Adobe Acrobat or the freeware CutePDF PDF driver on PC platforms, when a PDF image is created on IBM i, it must be directed to a file on the IFS. IBM i PDF support does not prompt you for a file name at runtime (at least I haven't had that experience) but instead requires that the file name be specified before the print file is open. It is remarkable how well this logic integrates with IBM i applications. If you're producing an A/R ledger report for example, the name LEDGER.PDF would be a good file name. Perhaps storing it in a folder named /home/apps/ar might be desirable. So the name you would identify to the OVRPRTF command would be: '/home/apps/ar/ledger.pdf'.

Magic PDF Command

The CL command and interfaces on v6r1 and later needed to create a PDF are so simple, it makes you wonder why IBM didn't implement it sooner. (That's a left-handed complement, by the way.) To create a PDF instead of a SPOOL file, add the following 3 parameters to an OVRPRTF command:

Parameter Definitions:

DEVTYPE - This is the device type. This parameter must always be set to DEVTYPE(*AFPDS) to create a PDF file.

WSCST - This is the Workstation customizing object. This parameter must always be set to WSCST(*PDF) to create a PDF file.

TOSTMF - The path and optional PDF file name to be created. You must specify an IFS path, but the file name may be auto-generated.

The DEVTYPE and WSCST are always the same, but the TOSTMF parameter is the one that varies based on the situation. Specify either a fully qualified PDF file name (IFS path and the PDF file name) or specify only an IFS path/ When an IFS path is specified and a PDF file name is not included, the system will generate a file name for you. Here are two examples that illustrate these options.


 In this example the file ledger.pdf in the /home/cozzi directory will be created from the output written to the QPRINT file.


In this example, the specific PDF name on the IFS is not specified. In this situation, the system generates a name using the Print file name and a random 6-letter pattern.


Where xxyyzz is a random character pattern that insures unique file naming.

To print source member from within PDM to a PDF file, I issued the following OVRPRTF command to the QPSUPRTF print file and then used PDM option 6 (print) on 3 of my source members.


This produced the following 3 PDF files in the /home/cozzi directory on my IFS:


To make the auto-generating names more clear, I have replicated them below, this time, separating the prefix used in the print file name from the unique identifier generated by the PDF driver:

QPSUPRTF 2h8Vtc.pdf

From this you can see that when the PDF driver is asked to generate a PDF file name, it uses  the print file name as a prefix, followed by 6 random characters, and then finally the '.PDF' file extension.


In addition to using OVRPRTF to direct output to a PDF file, you can permanently assign PDF output to an existing print file. Simply issue a CHGPRTF (Change Print File) command and specify the same values as used on the OVRPRTF command. If you are creating a new print file, you can easily specify PDF output by specifying the DEVTYPE, WSCST and TOSTMF parameters on the CRTPRTF (Create Print File) command.

Remember, if you are running IBM i v6r1, make sure you have 5661-TS1 installed. If you're running v7r1 or later, make sure you have 5770-TS1 and it should just work. You will need *BASE and Option 1 of the licensed program.

That's all there is to it.



For years IBM has been adding questionable interfaces to the system; mostly due to the fact that the System/38 Generation of Engineers retired or moved on to other things, but some are still there. This implementation of PDF support is very promising and surprised me; but in a good way. If you look at the great work IBM did here and in rewriting the CPYTOIMPF command back on V5R1 or V5R2 (replacing the previous, unusable implementation) it looks like IBM Rochester is really putting worthwhile efforts into this platform, and that's a good thing.

Call Me

If you have any questions on any area of RPG IV, Web development, API, C/C++ or anything else IBM i development related (except subfiles, data areas and RPGII/III because I don't care about that stuff) write a comment to me on our website and I may include a response in an upcoming issue of RPG Report.

You can subscribe to RPG Report (we call it "follow") by visiting RPG Report on midrangeNews.com and click the FOLLOW link in the table of contents on that page. To unsubscribe, simply click that same link. You must be signed up for, and signed in to midrangeNews.com to start following RPG Report.

-Bob Cozzi
Follow Bob on Twitter

Return to midrangenews.com home page.
Sort Ascend | Descend


(Sign in to Post a Comment)
Posted by: LarryTaylor
Premium member *
Comment on: RPG Report - 15 Feb 2011
Posted: 8 years 7 months 18 days 21 hours 20 minutes ago
Edited: Thu, 18 Aug, 2011 at 12:14:46 (3151 days ago)

Is there an RPGLE way to respool the PDF after it is created on the IFS?

I'm currently using TAATOOL DUPSPLF and infoPrint Server to create purchase order PDF's and copy them to 'LiveLink' for archive. The PO's are somtimes printed for international vendors in order to include additional documents before email. 

Thanks for all your help.

Posted by: bobcozzi
Site Admin ****
Comment on: RPG Report - 15 Feb 2011
Posted: 8 years 7 months 18 days 18 hours 37 minutes ago
Edited: Wed, 05 Feb, 2014 at 17:11:54 (2249 days ago)

You can copy a PDF on the IFS to a connected PDF-direct printer (most new and almost all network printers support PDF direct). Use FTP to do this, or use the CPYFRMPDF command found in COZTOOLS.

Posted by: LarryTaylor
Premium member *
Comment on: RPG Report - 15 Feb 2011
Posted: 8 years 7 months 17 days 19 hours 21 minutes ago
Edited: Thu, 18 Aug, 2011 at 12:14:46 (3151 days ago)

Thank you for your quick response.

TAATOOL DUPSPLF leaves a copy of the spooled file in the outq and we can print it.

I was hoping to use the new OVRPRTF options to create the PDF and print it.
I've defined a *USERASCII device and tried reading the PDF and writing to a couple of our network printers, but our printers do not convert the output to a readable format.

I also tried QSHELL (not really knowing how to use it) again the output was not in a readable format.

Thanks again. 

Posted by: khodges3tampabayrrcom
Premium member *
Comment on: RPG Report - 15 Feb 2011
Posted: 8 years 7 months 15 days 21 hours 22 minutes ago

There are a number of ways to do it.  Here are some.

IBM provides the PDF subsystem with it's infoprint server product. 

Scott Klement has a PRTSTMF command (I haven't used it).

You can ftp the file to the printer.

Posted by: LarryTaylor
Premium member *
Comment on: RPG Report - 15 Feb 2011
Posted: 8 years 6 months 8 days 21 hours 10 minutes ago

Thank you for your reply.

Update: We do use Infoprint Server, but would like to just use the OVRPRTF, and print directly from the IFS.

I can FTP a PDF from the IFS to our InfoPrint printers.

For the Xerox WorkCentre's (having issues with FTP )I had to :





this created a spool file of the PDF in my outq - printed just fine on the Xerox.


I also tried Scott's PRTSTMF - couldn't get it to work, probably something I don't understand completely.  

None of the above worked on our HP8150's.




Posted by: bobcozzi
Site Admin ****
Comment on: RPG Report - 15 Feb 2011
Posted: 4 years 8 months 14 days 17 hours 23 minutes ago

Larry, when you copy to a PDF you can physically copy the SPOOL file to PDF on the IFS, leaving the original SPOOL file in place. To print it, it needs to be binary-copied to the printer. Our CPYFRMPDF command in COZTOOLS does this binary copy. Your FTP steps seem to be fine.  Also when using Scott's code, remember while he is a very, very smart guy and produces some extremely clever stuff for IBM i users, I have often (always?) found that his code works when it is not changed from his implementation--otherwise it tends to be a challenge to get a customized version working. That's why I wrote iSockets, because I tried using his HTTP stuff, and whil it was very cool and very complete, it had to be installed in his library or it just wouldn't work (at least I couldn't get it to work). So use his stuff as examples of what can be done, and then write your own stuff for your own needs.