by I did not think it possible, but we have a number of records that have
the same value for the IDENTITY column in tables, which are
replicated!
I do not understand how this has happened because the situation should
have been treated as a conflict, which I have seen before, resulting
in data-loss.
Each table is defined with an IDENTITY column set to be
auto-incrementing with a default seed and increment of 1,1
respectively.
The server is SQL 7 (SP3) and the laptops with the replicated data are
running MSDE (SP3), Windows 2000 (SP2), MDAC 6.1, anonymous merge
replication over a RAS server via modem.
The replicated tables have dynamic filters and are set to use the
default resolver.
The records which have been duplicated have come from the Laptops.
The 'NOT FOR REPLICATION' option has NOT been set on the subscribers
(Laptops) and the IDENTITY seed and increment has NOT been changed
from the default of 1,1.
I would say that there is nothing out of the ordinary about our setup
Apart from this duplication, everthing else appears to be working
fine!
What is ironic is that for us, this situation, although faulty, is
actually better than having conflicts and data-loss!
The structure of 1 of the tables containing records with identical
IDENTITY column values:
CREATE TABLE [dbo].[marketing_matrix] (
[jcbf_id] [int] IDENTITY (1, 1) NOT NULL ,
[account_no] [varchar] (6) NULL ,
[xyear] [nvarchar] (4) NULL ,
[xmonth] [varchar] (3) NULL ,
[equipment] [varchar] (50) NULL ,
[unit_value] [varchar] (2) NULL ,
[COMMENTS] [varchar] (1000) NULL ,
[Returned_On] [datetime] NULL ,
[action_date] [datetime] NULL ,
[action_by] [varchar] (4) NULL ,
[complete_date] [datetime] NULL ,
[source] [varchar] (32) NULL ,
[rowguid] uniqueidentifier ROWGUIDCOL NOT NULL
) ON [PRIMARY]
Any takers?
Cheers
Mike Bredbury.