If your data set is very large, then instead of having the app download all rows on each synchronisation, you can implement incremental row updates based on a timestamp and deletion status.
This will result in only the changed/deleted rows being downloaded by the app after it’s first initial synchronisation, which should improve performance and reduce mobile data usage for workers in the field.
To enable, add the following columns to your SyncData tables (column names must match exactly and should be added to the end of the table):
This column contains the date-time value that the record was last updated and is used as a check when retrieving data to prevent retrieving old rows.
A boolean column that represents whether or not this row has been deleted.
As soon as you add a _lastUpdated column, then Sync will assume you want to run in incremental update mode and will use the _lastUpdated values as a row filtering mechanism, only returning rows that have _lastUpdated greater than the lastupdated parameter value received from the app's REST request to the Sync web service.
If you add a _lastUpdated column, ensure it is a not null column of DateTime data type.
All values must be in UTC timezone, and it is your responsibility to handle the updating of this _lastUpdated column whenever you make changes to the rows in that table.
e.g. if you add or update a row, then make sure _lastUpdated for that row is always set using the SQL Server getutcdate() function.
Additionally, if you are going to be deleting rows while using the _lastUpdated facility, then you must add a _deleted column which is not null and of BIT data type.
When you want to apply a deletion of a row from the table, instead of using the DELETE sql command, rather update the target row's _deleted to 1 (true) and set the _lastUpdated using getutcdate().
This is important because if you were to just delete rows from the table, then the app will never know about this and the row may persist for a period on the app going forward.
When implementing incremental updates, we strongly recommend that you try this out first in a testing account with a test install of Sync to see if it works for your specific scenarios.
Assuming it works well in your testing, then you should be able to simply reuse the same SQL scripts to apply incremental updating columns to your production environments in turn.