In article <[EMAIL PROTECTED]>, "Lonnie Princehouse" <[EMAIL PROTECTED]> wrote:
> Pickling is the Python term for serialization. See > http://en.wikipedia.org/wiki/Serialization > > Suppose you want to save a Python object "x" to a file... > > output_file = open('my_pickle', 'wb') # open a file > > import pickle > pickle.dump(x, output_file) # write x to the file > output_file.close() > > ... and to restore x from the file: > > input_file = open('my_pickle','rb') > x = pickle.load(input_file) > input_file.close() I used to use pickles a lot for making scripts start up faster by cacheing intermediate results. On startup, I had to read and parse a bunch of large text files and build a complicated in-memory database out of them. That took something like 10 seconds. However, the text files very rarely changed. To save startup time, I read the files in once, and pickled the database in a file. On subsequent runs, I'd just read in the pickle, which took a fraction of a second. -- http://mail.python.org/mailman/listinfo/python-list
