as i mentioned before i am still trying to use swish-e with the indexfiles
sitting on cd-rom. Therefore i am still trying to minimze file access.
The search in the btree itself is fast as lightning, but then it comes to
reading the worddata.
I found out that during a wildcard search in my indexfiles the worddata file
has to be opened about 16.000 times.
Therefore i am thinking about storing the worddata also in the btree file.
(replace the wordID with the worddata). I know that the btree file will get
a lot larger this way, but when i found the word the filepointer is already
in position to read the data.
Now i am trying to figure out what to change in the btree.c code to make
this work. Where i get lost are the BTREE_KeyData, BTREE_KeyDataOffset, and
Could someone point me into the right direction to make this work with a
char buffer (worddata) instead of an integer value (wordID)?
Received on Tue Mar 1 04:19:05 2005