To implement scripted upload, you create stored procedures that define the upload by returning result sets that contain the rows to update, insert, or delete on the consolidated database.
When the stored procedures are called, a temporary table called #hook_dict is created that has two columns: name and value. The table is used to pass name-value pairs to your stored procedures. Your stored procedures can retrieve useful information from this table.
The following name-value pairs are defined:
|timestamp as string||The time up to which all changes on the remote database have been uploaded. Your upload should only reflect operations that occur after this time.|
raw start progress
|64-bit unsigned integer||The start progress expressed as an unsigned integer.|
|timestamp as string||The end of the upload period. Your upload should only reflect operations that occur before this time.|
|raw end progress||64-bit unsigned integer||The end progress expressed as an unsigned integer.|
generating download exclusion list
|true|false||True if the synchronization is download-only or file-based. In those cases no upload is sent, and the download is not applied if it affects any row selected by a scripted upload stored procedure. (This ensures that changes made at the remote that need to be uploaded will not be overwritten by the download.)|
|publication name||The publications being synchronized, where n is an integer. The numbering of n starts at zero.|
|script version||version name||The MobiLink script version to be used for the synchronization.|
|MobiLink user||MobiLink user name||The MobiLink user for which you are synchronizing.|
See #hook_dict table.
Custom progress values in scripted upload
Defining stored procedures for inserts
Defining stored procedures for deletes
Defining stored procedures for updates
Creating publications for scripted upload