| Define unique index [message #625327] |
Mon, 06 October 2014 02:24  |
Tlg13team
Messages: 100
Registered: June 2008
Location: MGL
|
Senior Member |
|
|
Hello,
I try to implement database replication from Oracle to SQL Server.
My understanding all kind of log based replication approach requiring a unique index on source table.
When I update or delete a record on source table, replication software can found that record by unique index then made change on target table. If source table without unique index, every update and delete changes require full scan full scan on target table then impact database performance!!!
My source table without unique index! When I define a unique index, it's requiring 10 more column combination. It's impact data load performance.
I added a column on the table then populated by sys_uid() function then define unique index that column but it's still impacted data load performance.
How to solve this issue?
|
|
|
|
| Re: Define unique index [message #625334 is a reply to message #625327] |
Mon, 06 October 2014 03:10  |
John Watson
Messages: 8964
Registered: January 2010
Location: Global Village
|
Senior Member |
|
|
|
It is the target table that requires an index, not the source. The only change you need on the source is to ensure that the key column(s) is always logged, which you do by adding a supplemental log group.
|
|
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
