Identifying and publishing a PDF rendition to FTP server using FTP outbound and automatic activity

, All

In continuation to the thread , (which I sort of hijacked) , I am creating a new discussion

Scenario - Users import Pdf documents or word documents in the repository, ADTS creates PDF rendition of the office documents. Requirement is to publish PDF rendition of office documents or if the document format is PDF, publish the PDF document

There is already a solution in place that starts a workflow , code in first auto activity checks if original format of document is PDF, set a process variable to True, if rendition exists - set second process variable to True . The transition condition defines that if first variable is set to True, move to FTP Outbound Activity say A(it sets the File data to original document data), if second variable is True move to FTP Outbound Activity B (it sets the File data to Rendition[1] data); else send to error activity( that sends an email)

Problem - It is observed that many times this workflow is moving office documents to FTP server. This needs to be fixed so a correction needs to be done

Analysis - After analyzing it was observed that for faulty documents - FTP activity is getting office document in Rendition[1]/data in FTP Outbound Activity B but same solution is working other times ( so its intermittent issue)

After more analysis, it was observed that for faulty documents , when I click View Renditions - it shows PDF document first and office document second in UI and even while issuing query - select * from dmr_content where parent_id ='<<document id>>'

Conclusion - It seems workflow is designed assuming that rendition[1] will always have Pdf rendition but that is not the case

Fix being considered -

1) remove workflow totally and write the logic in a method (including FTP code) and call this method from webtop

2) call a method from web top - if the original document is PDF, start workflow on it; if it has renditions, create a temporary object, attach PDF rendition's content on that object and start workflow on it ( making sure workflow is always started on a pdf document and write some cleanup job to delete these temporary objects)

3) understand how rendition[1] gets the content and accordingly modify first workflow auto method and add one more FTP Outbound Activity C (that sends Rendition[2]/data)

Questions for #3 - is it always the first content returned by the dql ; can we rely on this logic ? if not how can we identify in original workflow's first activity which dmd_content belongs to rendition[1], which belongs to rendition[2]

Please note - in 4 or 5 cases in production we have observed two pdf rendition. probably I may skip that scenario for now

Shoeb,

Can you help how you set process variables with the values that exist for rendition[1] and [2] and suggest your views on above approaches

Kind Regards,

Ashish Srivastava

Find more posts tagged with

Note:

If there are Accepted Answers, those will be shown by default. You can switch to 'All Replies' by selecting the tab below.

Accepted answers

AshishSrivastava

Thanks Shoeb for your time. I was working with EMC Support also and BPM Enginnering team replied with the way rendition values are mapped to dmr_content values by Content Server. Here it is for everyone's reference

It is based on the IDfCollection object return by IDfSysObject.getRenditions API.

We iterate this collection and first value will be rendition[1], second rendition[2], third rendition[3] .

You can take a look at the rendition APIs of IDfSysObject to set the renditions in particular order.

I can use above details in first automatic activity and set the rendition with pdf format as rendition[1]

I can add one more activity that publishes rendition[2]. In auto activity , I can find if PDF rendition is at rendition[1] or rendition[2] and set the transition condition accordingly to go to correct activity

Hope this will help others also in case they have the same requirement to publish FTP rendition

Kind Regards,

Ashish Srivastava

All comments

Shoeb Haque

In your custom method, run a query to determine the primary content type (word or PDF) and also determine if the sysObject has a PDF rendition by querying the dmr_content for full_format = pdf. Once you both of these, you can do whatever flow you want in the method transition. i.e if PDF rendition does not exist, do nothing, else, if PDF rendition exists, use this in the mapping.

AshishSrivastava

Shoeb,

This is already is place and routing is also happening to correct FTP activity.

The problem is that even if document has a pdf rendition and it routes the the FTP Outbound activity that has to FTP the rendition - FTP activity assumes that rendition[1] is the PDF content which is not always true

Sometimes Rendition[1] is returning PDF and sometimes office document thus the issue

I checked with EMC support, few minutes back I got the below reply from BPM engineering

"If you have multiple renditions and you want to export a
particular rendition to FTP then you need to specify the index number of the
rendition you want to export otherwise how would BPM know which rendition to
export.

In this case, you are exporting the rendition which is at
rendition[1] from activity FTP Rendition.

This index is not fixed by the BPM. It has been fixed by the
customer expecting rendition[1] is always PDF.

Customer need to ensure that rendition[1] is always PDF. This is not a bug. work around is to fix the PDF rendition index
to 1 by custom java code. Looks like you will have to modify the custom
code to check and make sure pdf rendition fix to [1].

Even if I have to write a custom logic, I need to have a confirmation which dmr_content belongs to rendition[1] and which belongs to rendition[2] ( the backend query that Content Server uses for this could be anything. Say if it has a order by clause, the sequence of returned results will be opposite PDF first and then office format etc)

This is why I thought to go for Option 2 as a fix but if questions related to Option 3 can be answered, I can just fix the workflow code rather than changing the entire design

Kind Regards,

Ashish Srivastava

Shoeb Haque

I had the same requirement as yours and basically all I did was have a custom method before the FTP activity. In this custom method, I was setting a few process variables based on the query against the dmr_content etc.Then in the transition step of this custom method, I was checking for these Process Variables and based on the value forwarding it to the right FTP activity.

In your case, you can get the index of the dmr_content where format=pdf and use that. If index=0, ftp the primary content, if index=1, ftp the secondary one.....

AshishSrivastava

Thanks Shoeb, how can I get index from dmr_content ; can you provide the DQL for the same ?

Thank you !

Shoeb Haque

something like this:

DQL>select count(*) from dmr_content where any parent_id = '<sysObj id>' and full_format = 'pdf' and rendition= 1

This will tell you if the primary word document has a PDF rendition or not. You can then use this for transition as well.

You may have to play around with this query if this doesn't work completely.

AshishSrivastava

It is based on the IDfCollection object return by IDfSysObject.getRenditions API.

We iterate this collection and first value will be rendition[1], second rendition[2], third rendition[3] .

You can take a look at the rendition APIs of IDfSysObject to set the renditions in particular order.

I can use above details in first automatic activity and set the rendition with pdf format as rendition[1]

Hope this will help others also in case they have the same requirement to publish FTP rendition

Kind Regards,

Ashish Srivastava