Conditionally setting MERGE-BY-FIELD on a data source Cancel The title field is required!

1. Simon Prinsloo

   

   11 posts
   Registered:
   20 Feb 2019
   12 Mar 2020

   Link to this post

    Hello
   
   We have a condition where two different automated processes sometimes tries to update the same record within milliseconds from each other a few times a day.
   
   Both read the same version of the record from the database via dataset models. Each process then proceeds to update the record, but not the same field.
   Either one will succeed and the other will get the "Record changed by another user" error.
   
   Setting the MERGE-BY-FIELD property for these two cases will solve the problem, but I cannot leave it on all the time because the table is updated very frequently and I am concerned about the performance impact it would have on the 99.9% of the other updates where this is not needed.
   
   I am therefore looking for ideas about the best place to set this property and especially an effective way of signalling that I am busy with these specific two transactions, without creating a messy structure.
   
2. Mike Fechner

   

   342 posts
   Registered:
   14 Sep 2016
   13 Mar 2020 in reply to Simon Prinsloo

   Link to this post

   This is a very broad question.
   
   There are many places where you can set that property in the Data Access class.
   
   The final possibility is probably in the override to the SaveRowChanges method. Which conditions to use depend a lot on your use case.
   
   An alternative might also be to override SaveRowChances and handle the writing from the TT to the DD yourself, instead of relying on the SAVE-ROW-CHANGES method.