Reltio Connect

 View Only
Expand all | Collapse all

Starting a Potential Match Workflow from API without Duplicates - generateFromQuery

Jump to Best Answer
  • 1.  Starting a Potential Match Workflow from API without Duplicates - generateFromQuery

    Posted 10-22-2021 11:56
    Hello folks,

    My investigation and experimentation with Workflow is going quite well and I have learned a ton.  Thanks for being there for me as I go through this process.

    The latest I've done is to use the processInstances/_generateFromQuery API call in order to automate the assignment of potential matches to workflow reviewers (see screen example below).  I include a query for search string that will bring back a small number...since I am just experimenting and don't want to assign too many right now.  The api executes and indeed it does assign these tasks.  However, it assigns a separate task for each potential match.  This appears to me to be duplicates in the sense that once I open one party profile I will have access to merge all of the other parties so I don't need that other task entries filling up my inbox.  It sort of makes sense that it would do this...since the search string if I ran it to retrieve the entities alone would do the same.  When looking in the Filtering documentation there's nothing obvious to me that could eliminate duplicate in the output.  I'm not sure what the best practices are surrounding this approach...but this is where my experimenting is leading me.

    Another question that I have is regarding this Note  below in the above documentation.  Does this mean that I should not start the processInstances api call many times or that it shouldn't have a query that will create too many individual workflow tasks to be assigned to reviewers?  My thinking was that I could have a process at the end of our load cycle that could execute a number of api calls with different search strings to assign potential matches meeting a certain criteria to reviewers.  Again, do let me know if this fits in with best practices.  The reason why i'm thinking to automate things is to make things easier on our data stewards - although we have "data stewards" they are not full time in this role and so pushing out matches meeting specific criteria automatically  sounds like it could be really useful - I'm just not sure if the number meeting a specific criteria would stay under 100....and what might happen if it didn't.

    Note: The _generateFromQuery request has a HTTP session timeout limitation. Therefore, Reltio recommends not to start more than 100 processes at the same time.


     "processType":"potentialMatchReviewTest1",
     "searchString":"(equals(type,'configuration/entityTypes/Individual') and (gt(matches,'2') and equals(attributes.LastName,'KVERN'))"
     
     }




    ------------------------------
    Michelle Popovits
    Empire Life
    ------------------------------


  • 2.  RE: Starting a Potential Match Workflow from API without Duplicates - generateFromQuery
    Best Answer

    Reltio Employee
    Posted 10-25-2021 09:02
    Hello Michelle,

    > the sense that once I open one party profile I will have access to merge all of the other parties
    it is not always true. Metadata Security and custom workflow implementations might be very complex and flexible - depending on entities data each workflow task can be assigned to users (Data Stewards) with different permissions.
    To solve your concern about duplicate tasks - if there is already a Potential Match Review (PMR) workflow process started for a couple of entities it cannot be started for them again.

    >  Does this mean that I should not start the processInstances api call many times or that it shouldn't have a query that will create too many individual workflow tasks to be assigned to reviewers?
    It means the search query should not return too many entities because it is leading to starting too many workflow processes. Currently, this API works in synchronous mode that guarantees all entities were processed and workflow processes were started. If the request for starting PMR workflow processes is canceled by HTTP timeout it is not guaranteed.

    >  I'm just not sure if the number meeting a specific criteria would stay under 100
    you can always do a search request to get the number of entities for the criteria before starting PMR workflow.

    Best regards,
    Yury


    ------------------------------
    Yury Timofeev
    ------------------------------



  • 3.  RE: Starting a Potential Match Workflow from API without Duplicates - generateFromQuery

    Posted 10-25-2021 14:55
    Thanks Yury!

    ------------------------------
    Michelle Popovits
    Empire Life
    ------------------------------