Creating Primary DB and Secondary DB

Michael Cahill · *nix forums Guru Wannabe Joined: 26 May 2005 Posts: 219

Searching with DB_SET_RANGE uses the b-tree structure, it isn't a
linear scan.

Michael.

Wendeline · *nix forums beginner Joined: 26 Mar 2005 Posts: 17

Hi,

Using DB Cursor and set the flag to DB_SET_RANGE, does it mean the
cursor will actually travsers sequentially all the record untill it
finds the right id or it use b-tree algorithm to traverse ? If it is
the first case, then having three b-tree databases should be better
than having two databases? Am I right?

Florian Weimer · *nix forums Guru Joined: 19 Feb 2005 Posts: 418

Wendeline · *nix forums beginner Joined: 26 Mar 2005 Posts: 17

Hi, I need a database for storing ids and strings, showing the
following behaviors:
i) Both id and string can have duplicates
ii) id+string is unique.
iii) id->string mapping and string->id mapping should be available.

I tried to create a primary DB with id as the primary key and associate
it with a secondary index DB with string as secondary key, then I
realized that primary key may not be configured with duplicates. So I
guess I have to create a primary DB with id+string as the key, and
data could be the same value as key or even null since all the
information are included in the primary key. Then I create two
secondary DBs to associated with the primary DB, one with id as
secondary key, the other with string as secondary key...Ended up I have
to create 3 DBs to have the above 3 behaviors working.

Could there be a simplier way to do this?

Google