we discovered a situation where
a target table for cdc has no unique index nor unique constraint
but for the moment we do not have any duplicates
in cdc subscription the key was specified as automatic - but there is none
will cdc use the entire row as index ?
we can change the subs to specify the key and select all columns. but there is anyhow a potential problem if nullable columns are allowed
we tried to create an unique index on table, but this table is very large and we get out of swap space.
we need to ask to change the machine to get more.
question : will cdc use the entire row as index ? and fail if dupl row is encountered I presume
is cdc not checking this when you setup the subs that nothing is unique on table ? is it allowed ?
------------------------------
Thanks for all answers
Best Regards,
Guy Przytula
------------------------------
#DataIntegration#DataReplication