prefix matching fuzzy algorithms; working!!!!!


Jesse op den Brouw (jesse@crytonII.st.hhs.nl)
Tue, 31 Mar 1998 23:03:21 +0200


Hello Esa, Andrew

I've got it working!!!! Instead of gdbm I've installed berkeley DB
version 2.3.16 which is different from 1.85.

Compiled on Linux RH 4.2 / P75 / 16MB RAM / libc 5.3.12

What was the problem?

Well, in gdbm, fetch routine allocates space that the user has to free.
Berk-DB uses some automatic alloc scheme; the user must not free
the allocated space or you wind up in sigsegv's.....
So I deleted the free's in Get, Get_Next etc....

Esa, can you check if this free should be done in Berk-DB 1.85?

Then I discovered I made a mess of my libc.* 's so in the linking fase
it went beserk.

The prefix is working well (at least for now) on a 60 MB database with
7400 docs an 78000 different words.

Two questions:

1) Are the databases multiple key?

2) Is the performance of a btree database (always) better than that of
   a flat base (gdbm?)?

Greetings from holland

--jesse
---------------------------------------------------------------------
J. op den Brouw Johanna Westerdijkplein 75
Haagse Hogeschool 2521 EN DEN HAAG
Faculty of Engineering Netherlands
Electrical Engineering +31 70 4458936
-------------------- J.E.J.opdenBrouw@st.hhs.nl ---------------------

Linux - because reboots are for hardware changes



This archive was generated by hypermail 2.0b3 on Sat Jan 02 1999 - 16:25:51 PST