summaryrefslogtreecommitdiffstats
path: root/Doc/lib/libshelve.tex
blob: 3e2bef930371070be8bfb72ea5b93199757998ca (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
\section{Built-in module \sectcode{shelve}}
\stmodindex{shelve}
\stmodindex{pickle}
\bimodindex{dbm}

A ``shelf'' is a persistent, dictionary-like object.  The difference
with ``dbm'' databases is that the values (not the keys!) in a shelf
can be essentially arbitrary Python objects --- anything that the
\code{pickle} module can handle.  This includes most class instances,
recursive data types, and objects containing lots of shared
sub-objects.  The keys are ordinary strings.

To summarize the interface (\code{key} is a string, \code{data} is an
arbitrary object):

\begin{verbatim}
import shelve

d = shelve.open(filename) # open, with (g)dbm filename -- no suffix

d[key] = data   # store data at key (overwrites old data if
                # using an existing key)
data = d[key]   # retrieve data at key (raise KeyError if no
                # such key)
del d[key]      # delete data stored at key (raises KeyError
                # if no such key)
flag = d.has_key(key)   # true if the key exists
list = d.keys() # a list of all existing keys (slow!)

d.close()       # close it
\end{verbatim}

Dependent on the implementation, closing a persistent dictionary may
or may not be necessary to flush changes to disk.

Note: \code{shelve} does not support {\em concurrent} access to
shelved objects.  Two programs should not try to simultaneously access
the same shelf.