Persistent to file dictionary using Python module shelve

For the sake of comparing tools, here is the same code written with the well known zodb module instead of shelve.

''' zodb_test.py
Use third party module ZODB (zope object database)
to create a 'persistent to file' dictionary

Python 2 and 3 create files
    'phonebook.fs'
    'phonebook.fs.index'
    'phonebook.fs.lock'
    'phonebook.fs.tmp'
The files are not compatible between Python 2 and 3.

Tested with Python27 and Python33 by Gribouillis 2013-Sep-05
'''

# There are a few more necessary imports with zodb
from ZODB.FileStorage import FileStorage
from ZODB.DB import DB
import transaction
from BTrees.OOBTree import OOBTree

# name zodb data file
data_file = 'phonebook.fs'

# open existing zodb storage file or create new file
storage = FileStorage(data_file)
db = DB(storage)
connection = db.open()
sh = connection.root() # <- our starting point, the database root object.

# create a persistent dictionary (technically a b-tree)
phonebook = sh["myphone_dict"] = OOBTree()

# update from an ordinary python dictionary
phonebook.update({
'Andrew Parson': '880-6336',
'Emily Everett': '678-4346', 
'Peter Power': '765-8344',
'Lewis Lame': '112-2345'
})

# do one more dictionary entry ...
phonebook['Ginger Roger'] = '123-4567'

# save to database before you close
transaction.commit()

# close when done. A few more calls than with shelves.
connection.close()
db.close()
storage.close()

# testing ...
# open the zodb file again for further work
# remember the access key is "myphone_dict"
storage = FileStorage(data_file)
db = DB(storage)
connection = db.open()
sh = connection.root()

phonebook2 = sh["myphone_dict"]

# can add another dictionary entry
phonebook2['Larry Lark'] = '923-4568'

# save before you close
transaction.commit()

# close when done
connection.close()
db.close()
storage.close()

print('-'*30)

# test the contents ...
# retrieve the saved data from the file
storage = FileStorage(data_file)
db = DB(storage)
connection = db.open()
sh = connection.root()

phonebook3 = sh["myphone_dict"]

# print the contents
for name, phone in sorted(phonebook3.items()):
    print( "%-20s  %s" % (name, phone) )

print('-'*30)

# close when done
connection.close()
db.close()
storage.close()

''' result ...
------------------------------
Andrew Parson         880-6336
Emily Everett         678-4346
Ginger Roger          123-4567
Larry Lark            923-4568
Lewis Lame            112-2345
Peter Power           765-8344
------------------------------
'''

Next step: comparative benchmarks :). Notice that zodb does not load the whole dictionary when it is accessed, because we chose the OOBTree class. It is suitable for very large data sets.

For small persistent dictionnaries, there is also a very nice implementation by Raymond Hettinger which let the user choose between json, csv or pickle formats !