模仿关系数据库的Python数据结构

Question

I would like to store a lot of instances of some data in python. Each record has the following fields: username, address, salary etc...

The username should be unique. I do a lot of searches.

Currently I am using a list of dictionaries, but when I insert a new item I iterate the list and check the username of each dictionary, which is O(n). Searching is O(n) too. How could I achieve that there is an index on usernames, and make the search time O(logn)?

Answer 1

Why not use a dictionary of dicts?

The top-level dictionary could be addressed by the username and the other dicts can be the same you have now.

A dictionary can also be scanned threw (d.values()) -- the only disadvantage is, that you can't depend on the ordering.

Of course that is not a DB behavior, but in most cases good enough and very fast -- the access via dict is O(1).

Of course you could use sqlite -- but when you just want to access via username and scan the entries, you are much faster (both in development- and runtime speed).

Answer 2

You could alternatively actually use a relational database. The sqlite module allows for databases in memory:

import sqlite3
conn = sqlite3.connect(':memory:')
# ...
conn.close()