I possibly could; but, I not only create the tree and get my stats, I also fold up nodes based on their leaf count, and re-calculate the values for the number of bits needed per prefix, the number of internal nodes, and the number of prefixes stored at each level. I don't thik that the Berkeley DB gives me enough flexability to do this, though I may be wrong.
Oh and incase you're wondering, when I fold up the leaves, I do delete them, to try to get some space back. I also remove the weight from the leaves, as it is no longer needed. | [reply] |