3

I faced this issue, that in nearly 10 GB txt files I must search for specific words or expressions. And I must use B+tree for that.

I also face the limitation of using memory: this allowed to use only 15% of memory used by my main files for extra memory(for saving b+tree and indexes of course.)

And the timing must be the most efficient (seriously I didn't get that my self!) It has something to do with my hard sectors and clusters. For least access time to memory.

Can you advise me how to set indexes to have an efficient program?

Regards

Fezo
  • 183
  • 2
  • 14
  • can you be more specific on the platform (dbms, os, hw) and the development environment ? have you looked into the native doc management systems (e.g. oracle text)? a generally idea is to compute the [inverted index](http://en.wikipedia.org/wiki/Inverted_index) of the text files upon document insertion and store this inverted index in your database (many of which will use some kind of b(+)-trees for indexing anyway). in case you aren't backed by a rdbms, [this so post](http://stackoverflow.com/questions/1720738/looking-for-a-disk-based-b-tree-implementation-in-c-or-c) might help you. regards. – collapsar Jan 22 '13 at 08:41
  • Use rocksdb or wiredtiger. – amirouche Oct 16 '20 at 21:42

0 Answers0