The index options ignore_dup_key,
ignore_dup_row, and allow_dup_row control what happens
when a duplicate key or duplicate row is created with insert or
update.
This table shows which option to use, based on the type of index.
Index Type
Options
Clustered
ignore_dup_row | allow_dup_row
Unique clustered
ignore_dup_key
Nonclustered
None
Unique nonclustered
None
ignore_dup_key Option
If you need to insert a duplicate value, you can use ignore_dup_key option with a unique index.
sorted_data Option
The sorted_data option of create index speeds index creation when the data in the table is already in sorted order, for example, when you have used bcp to copy sorted data into an empty table.
on segment_name Option
A nonclustered index can be created on a different segment than the data pages. The on segment_name clause specifies a database segment name on which the index is to be created.