在 memcached 中安全存储 Python 字符串和整数元组的最佳方法？

Question

So I have a webapp where I want to store current rankings of posts based on an algo every minute. To render the updated posts, I would like to avoid the db and get the data (eg post title, author, rank, etc.) from memcached (faster, right?).

One idea is (I'm using Django btw) for the view to convert a serialized string from memcached back to python (using pickle) and then rendering a template displaying the posts in the correct order. However, it may be unsafe because the user-submitted data (post title, etc.) may contain characters that may be parsed in an undesirable way.

1) Is the best solution to just clean all data when these objects are saved or is there another solution?

2) If cleaning the data is an option, what characters should be excluded in the various fields of my Post model?

3) Does pickle.dump require a file to write to, or can I just get a string as output?

Thanks guys.

An example of what needs to be stored:

((post.id, post.title, post.upvotes, post.author.username), (,,,)... )

(say 25 posts repeating like this)

Answer 1

Why not just let django.core.cache take care of that for you? I'm doing this from a django project that has memcached hooked up as the cache backend:

$ python manage.py shell
In [1]: from django.core.cache import cache

In [2]: cache.set('woot', (3, 'blah', 590))

In [3]: cache.get('woot')
Out[3]: (3, 'blah', 590)

It works for any python object that can be pickled.

Also, why not just cache the output, with HTML and everything? Django provides good per-view caching and template fragment caching.