Handle Concurrency

Question

Hi, 
 I have a simple process with one user input task to update the reference data / attributes in the DB. There are multiple users who can update the data.I need to handle a concurrency to update the attributes. At any point in time i need to allow only one user / one active process to update the attributes. 
 Are there any Appian functions to return active process details ? 
 
 Thx.

aloks0189 · Answer

Hi ravinderl352 you can use @Version annotation for those fields on which you want to have locking mechanism enabled. 
 For this you need to download the xsd of the corresponding CDT, define @Version annotation this will make sure to release the lock only when one transaction is completed. And in fact you can capture the update version to keep the track of, how many times this row have been modified. 
 Hope this will help you

aloks0189 · Answer

Hi Vinay 
 
You need to have a version comparison check before writing it to DB in order to avoid the process model node breaks. 
 
This is something which we can deal with. And there are various solution approaches for a particular use case. 
 
And the fact is, @Version is meant for handling the concurrency, if in every single case this lead a process to break, then Appian might not have provided the support for this annotation as part of jpa standards. 
 
Let's eloborate this point, When you want to assign a task to a group, that means concurrency is not your target. 
 
Here is an example: 
 
 A company wants to have a survey on like and dislike basis. Where they will have only one row on db which will have few columns including likeCount and dislikeCount and one task for each user has been assigned. Where the organisation is having thousands of employees, and every single users like or dislike will update the same row values again and again by increasing its count. 
 
Here in this scenario, assigning the task to a group is not the solution, instead we are trying to bypass the problem. 
 
In fact we should assign one task to each user. Hence here you need to handle the concurrency