0

I am going to be using merge replication in SQL 2012 for offline clients who can synchronise their data.

The merge replication process adds a guid to each table that is being replicated, but I am wanting to know how to add an additional int id as well?

For instance our clients ring up the support desk and they have to identify a particular record they don't want to have to deal with guids.

An identity column would be ideal, but how would this work with replication? Would each site have to have a different identity range?

peter
  • 13,009
  • 22
  • 82
  • 142
  • Yes, a different identity range would be one approach. However I am puzzled: what does the additional int id buy you? – Aaron Bertrand May 30 '12 at 21:30
  • If two people want to identify the row they don't want to have to read out a GUID over the phone. People will ring our support desk and say 'road 1452 has blah blah problem that I need to solve' – peter May 30 '12 at 21:45

1 Answers1

1

The answer is here,

https://dba.stackexchange.com/questions/18807/what-happens-when-identity-range-isexceeded

Basically each client is given two ranges. Once they have used up the first range and they sync they will be given a new range.

An error is thrown if an insert occurs in a table that has exceed both ranges. The only way this could happen is if a user inserts 2000 records before doing a sync. The range size can be increased if this is likely.

Community
  • 1
  • 1
peter
  • 13,009
  • 22
  • 82
  • 142