Thank you for answering the question! Like I said in my other post, the index creation is about 20% faster than the unique index creation in my latest tests of an index of two INTEGERs. 20% out of several hours is significant.
So index names are unique for the whole database? I was thinking they just had to be unique for the table. I'll double check my names... > > Is there some way I can create that unique index without > the overhead > > of uniqueness checking? I know all my triples are unique when used = > > together: I > > generated them that way. Or is there some other way I can > create that > > = index faster? > > > > The following technique is unsupported. It might change or > go away in a subsequent release. It might not work. > It might corrupt your database. Use with extreme caution. > > (1) CREATE INDEX primkey ON data(x,y,z) > (2) PRAMGA writable_schema=ON > (3) UPDATE sqlite_master > SET sql='CREATE UNIQUE INDEX primkey ON data(x,y,z)' > WHERE name='primkey' > (4) close and reopen the database connection > > The above technique is unsupported. It might not work. > It might corrupt your database. The overhead of the > uniqueness check is not that much so it is unlikely to help > much anyway. You have been warned. > > -- > D. Richard Hipp
