Location of Uploaded IPAs

IPA flows aren't stored in the file system as they are in PFI. They are uploaded as xml into the PfiFlowDefinition/PfiFlowVersion business classes.

By the way, you also won't find workunit logs on the server. The logs are saved into a CLOB field on the WorkUnit business class.

Good Luck with your transition.

They're in the Landmark database, not on the Landmark server. I've tried to look for them. Flow 'header' information is easily enough found....but the actual XML that is the flow is buried in one of the obscurely named tables.

If you're looking for a SQL query, here's something to start with (we're on DB2 SQL, so you may need to convert it if you're on MSSQL. Note that workunits (and all IPA data) are stored in the db in the appropriate application schema rather than in LOGAN:

select
pfd.pfiflowdefinition
,pfd.description as flowdescription
,pfd.currentflowversion
,case pfd.flowloglevel
when 1 then 'None'
when 2 then 'Work Unit Only'
when 3 then 'Activity Only'
when 4 then 'Work Unit and Activity'
end as flowloglevel
,pfv.comment as flowversioncomment
,pfv.uploaddate
,pfv.uploadby
,pfv_lob.flowxml
from
{dbSchema}.pfiflowdefinition as pfd
inner join {dbSchema}.pfiflowversion as pfv
on pfv.pfiflowdefinition = pfd.pfiflowdefinition
and pfv.pfiflowversion = pfd.currentflowversion
inner join {dbSchema}.s$pfifv as pfv_lob
on pfv_lob.uniqueid = pfv.uniqueid
where
pfd.pfiflowdefinition = {whatever}

Thanks! I thought when I was in IPA training class, there were lpd files out on Landmark. I noted the directory and looked in the same directory when I got back to work only to not find them.

By the way, for the workunit log, they are in S$PFIWU. You'll need to join to WorkUnit to look it up by number. However, it is really much easier to open the log in RichClient (opens in your text editor). However, huge logs are much more troublesome than they are on PFI because RichClient will run out of memory trying to open them....which pretty much sucks.

Regarding the files you saw in IPA training class, those may have been the "converted" files. They may have placed the PFI files on the file system, then ran the conversion program that makes them into IPA xml files. These files are then uploaded via Designer. Maybe that's what you're remembering? Just a thought.

Unless I am missing something the .lpd files are the ones you use/create within IPA Designer.  When you upload them to the server is when they get converted to the the .xml.  Therefore, your .lpd and .xml files should be equivalent.  You have the ability to put the .lpd files anywhere you want.

 

 

Sweet query, Woozy! Kudos!!

A related question for the IPA gurus:
Where is the UA approver name /RMID held in IPA-Landmark?
We had 2 queries that handled this in version 9.
The first DME queried LOGAN tables: Query string: PROD=LOGAN&FILE=WFUSERCAT&INDEX=WFTSET1&FIELD=WF-RM-ID&SELECT=TASK=Manager%26WORK-CAT-VALUE=002711620%26WORK-CATEGORY=Req%20Loc&OUT=CSV&DELIM=~
---
The 2nd took this result an looked up the Manager's name via a Resource Query:
)]]>

Hi Todd - IPA queries are created as xml files by Designer (on your IPDesigner workstation). When you use the conversion utility on PFI .lpd files they are cloned and a new .xml file in the IPA format with the same name is created. You then open the .xml file in Designer and upload it to the Landmark server where it is stored in the database as xml data. You can store the source .xml files anywhere you want to be used with designer, but they are not stored on the server file system that way by default.

Obviously, if your Landmark server is on Windows, and you do your development on that server rather than a separate Windows client, then the files may be stored on that machine, but that's a choice you make as a customer.

Hi jdever - That is a bit difficult question to answer. For IPA flows (I think this applies for LSF flows as well as TM flows) I think you have to drill from WorkUnit to PfiQueue to PfiQueueTask to PfiMetrics. This will tell you who took action, what action, and when. Then you'll have to determine the associated employee and query that info. This is ugly in Landmark. I've never taken the time to work it all out because it is so ugly, and it's generally easier to look at the workunit to find the status for a particularly flow. It would be nice to have a query to show us where stuff is held up, but we don't have one right now.