Hackers,
I'm starting to read the existing algorithms for btree index shrinking.
Right now I'm at 1996 SIGMOD proceedings, Zou and Salzberg "On-line
Reorganization of Sparsely-populated B+-trees".
What I want to know is how different from B+-trees are PostgreSQL
B-trees; I've read the README in src/backend/access/nbtree/, and it
indicates some areas in which they are different from B-Trees (Lehmann
and Yao's?). But I don't really know how B-Trees are different from
B+-Trees (is my ignorance starting to show?). Where can I read about
that?
Also, Tom said some time ago that there is some literature on the
concurrent page merging camp. I haven't been able to found anything
else than the proceedings I have right now... is there something else?
I'm not used to searching for this kind of things, and ACM won't let me
in (althought my university has a subscription, I can't get any papers
on SIGMOD).
Thank you,
--
Alvaro Herrera (<alvherre[a]atentus.com>)
"Un poeta es un mundo encerrado en un hombre" (Victor Hugo)