[sqlite] Implications of multi-column UNIQUE?

[Frank J. Iannarilli]

Hi,
Perhaps this is a question only drh can anwser, but here it is anyway.
Let's create a table with a multi-column UNIQUE constraint, e.g.:
   create table myTable (
     key integer primary key,
     col1 float not null,
     col2 ....
     ...
     colN type,
     UNIQUE (col1, col2, ... ColK)
   );
When there is a **multi-column** UNIQUE constraint, no **combination** of 
said column values can occur as duplicate-record (yes this works - I tried 
it).

In the online docs <http://www.sqlite.org/lang.html#createtable>, it states:
"The UNIQUE constraint causes an index to be created on the specified 
columns".

It's obvious what that means for *individual* column constraints, but what 
exactly occurs for the multi-column UNIQUE constraint?

What are the time-complexity implications of this, specifically for record 
insertions?  For instance, what if I have 10-20 columns in such a 
constraint?

Thanks!

Frank J. Iannarilli, [EMAIL PROTECTED]
Aerodyne Research, Inc., 45 Manning Rd., Billerica, MA 01821 USA
www.aerodyne.com/cosr/cosr.html
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]