Is pure Python a persistent key and value-based container (hash interface) with great file system support?

Question

I am looking for a (possibly) pure Python library for a persistent hash table (btree or b + tree, which would provide the following functions

• Support for large files (possibly in terabytes)
• Fairly fast and low memory (search for a balance of descent between speed and memory)
• Low cost management
• Reliability, i.e. does not corrupt the file when the contents are written through the file system.
• Finally, a pure Python implementation. I am fine if I have a C library, but I am looking for a cross-platform solution.

I reviewed solutions like redis, shelve, tokyo cabinet. The Tokyo office is impressive and has a Python binding when creating http://code.google.com/p/python-tokyocabinet/ , but its Windows port is work.

Thanks for the good suggestions. I am currently learning SQLite3 with Python. I have suggestions for using the database engine, but I'm more prone to meager and medium persistent implementations of the b + tree.

Answer 1

ZODB is really a powerful tool, but perhaps it is overkill.

Python: . , SQLite MySql, .

Answer 2

ZODB
http://pypi.python.org/pypi/ZODB3

,

Answer 3

.

• , .
• :
• -
• JOIN s, , , .

, dict -, . , .