Has anybody used the WB B-tree library?

Question

I stumbled across the WB on-disk B-tree library:

http://people.csail.mit.edu/jaffer/WB

It seems like it could be useful for my purposes (swapping data to disk during very large statistical calculations that do not fit in memory), but I was wondering how stable it is. Reading the manual, it seems worringly 'researchy' - there are sections labelled [NOT IMPLEMENTED] etc. But maybe the manual is just out-of-date.

So, is this library useable? Am I better off looking at Tokyo Cabinet, MemcacheDB, etc.?

By the way I am working in Java.

Answer 1

I have looked at the WB B-Tree Database, but SQLite might be a better fit. It handles extremely large datasets in a single file, and is a lightweight, fully-functional database.

http://www.sqlite.org/

Info on using SQLite with Java is here:

http://stackoverflow.com/questions/41233/java-and-sqlite

Answer 2

Yup, I gave it the good old college try in java. The jar file was easy to find as was the documentation. I think it was written in Scheme or something the likes and was translated to be usable in java.

The documentation speaks of functions that you ought to use but not what Objects they reside on. Sadly there is no java doc to help me out... There are no working examples and after 2 hours of trying I finally gave up. I found it not very useful at all.

I hope others have better luck using it.