IDENT_CURRENT('tableName') (include the single quotes) returns the current value of the identity for the given table. This value should be the last-assigned identity value used in the table. In other words, you will have a row with this identity value already in the table, unless that row has been deleted. The identity value that will be assigned on the next INSERT will be IDENT_CURRENT('tableName') + IDENT_INCR('tableName').
I don't recommend relying on this, though. If you pre-determine the next identity value this way, you're bound to end up in a situation where another process makes the insert that actually gets that ID before yours does, so your process ends up using the wrong ID value.
It's much better to make your insert first (even if you don't have all the data yet), and use SCOPE_IDENTITY() to get the actual ID assigned.
You might wonder why SCOPE_IDENTITY() is better than IDENT_CURRENT('tableName'). As the name implies, the former will give you that most recent identity value assigned within your current scope (your batch, your stored proc, whatever), whereas the latter will give you the most recent identity assigned on the table, by anyone. Even though you might call IDENT_CURRENT right after 'INSERT, it's still possible that someone else's INSERT occurs in between, and IDENT_CURRENT will give you the identity value that resulted from their insert instead of yours, whereas SCOPE_IDENTITY will always give you yours.
eg
SELECT IDENT_CURRENT('table_name') as CurrentIdentity,
IDENT_INCR('table_name') as Increment,
IDENT_CURRENT('table_name') + IDENT_INCR('table_name') as NextIndentity
Submit an article or tip