• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1:mod:`sqlite3` --- DB-API 2.0 interface for SQLite databases
2============================================================
3
4.. module:: sqlite3
5   :synopsis: A DB-API 2.0 implementation using SQLite 3.x.
6.. sectionauthor:: Gerhard Häring <gh@ghaering.de>
7
8
9.. versionadded:: 2.5
10
11SQLite is a C library that provides a lightweight disk-based database that
12doesn't require a separate server process and allows accessing the database
13using a nonstandard variant of the SQL query language. Some applications can use
14SQLite for internal data storage.  It's also possible to prototype an
15application using SQLite and then port the code to a larger database such as
16PostgreSQL or Oracle.
17
18The sqlite3 module was written by Gerhard Häring.  It provides a SQL interface
19compliant with the DB-API 2.0 specification described by :pep:`249`.
20
21To use the module, you must first create a :class:`Connection` object that
22represents the database.  Here the data will be stored in the
23:file:`example.db` file::
24
25   import sqlite3
26   conn = sqlite3.connect('example.db')
27
28You can also supply the special name ``:memory:`` to create a database in RAM.
29
30Once you have a :class:`Connection`, you can create a :class:`Cursor`  object
31and call its :meth:`~Cursor.execute` method to perform SQL commands::
32
33   c = conn.cursor()
34
35   # Create table
36   c.execute('''CREATE TABLE stocks
37                (date text, trans text, symbol text, qty real, price real)''')
38
39   # Insert a row of data
40   c.execute("INSERT INTO stocks VALUES ('2006-01-05','BUY','RHAT',100,35.14)")
41
42   # Save (commit) the changes
43   conn.commit()
44
45   # We can also close the connection if we are done with it.
46   # Just be sure any changes have been committed or they will be lost.
47   conn.close()
48
49The data you've saved is persistent and is available in subsequent sessions::
50
51   import sqlite3
52   conn = sqlite3.connect('example.db')
53   c = conn.cursor()
54
55Usually your SQL operations will need to use values from Python variables.  You
56shouldn't assemble your query using Python's string operations because doing so
57is insecure; it makes your program vulnerable to an SQL injection attack
58(see https://xkcd.com/327/ for humorous example of what can go wrong).
59
60Instead, use the DB-API's parameter substitution.  Put ``?`` as a placeholder
61wherever you want to use a value, and then provide a tuple of values as the
62second argument to the cursor's :meth:`~Cursor.execute` method.  (Other database
63modules may use a different placeholder, such as ``%s`` or ``:1``.) For
64example::
65
66   # Never do this -- insecure!
67   symbol = 'RHAT'
68   c.execute("SELECT * FROM stocks WHERE symbol = '%s'" % symbol)
69
70   # Do this instead
71   t = ('RHAT',)
72   c.execute('SELECT * FROM stocks WHERE symbol=?', t)
73   print c.fetchone()
74
75   # Larger example that inserts many records at a time
76   purchases = [('2006-03-28', 'BUY', 'IBM', 1000, 45.00),
77                ('2006-04-05', 'BUY', 'MSFT', 1000, 72.00),
78                ('2006-04-06', 'SELL', 'IBM', 500, 53.00),
79               ]
80   c.executemany('INSERT INTO stocks VALUES (?,?,?,?,?)', purchases)
81
82To retrieve data after executing a SELECT statement, you can either treat the
83cursor as an :term:`iterator`, call the cursor's :meth:`~Cursor.fetchone` method to
84retrieve a single matching row, or call :meth:`~Cursor.fetchall` to get a list of the
85matching rows.
86
87This example uses the iterator form::
88
89   >>> for row in c.execute('SELECT * FROM stocks ORDER BY price'):
90           print row
91
92   (u'2006-01-05', u'BUY', u'RHAT', 100, 35.14)
93   (u'2006-03-28', u'BUY', u'IBM', 1000, 45.0)
94   (u'2006-04-06', u'SELL', u'IBM', 500, 53.0)
95   (u'2006-04-05', u'BUY', u'MSFT', 1000, 72.0)
96
97
98.. seealso::
99
100   https://github.com/ghaering/pysqlite
101      The pysqlite web page -- sqlite3 is developed externally under the name
102      "pysqlite".
103
104   https://www.sqlite.org
105      The SQLite web page; the documentation describes the syntax and the
106      available data types for the supported SQL dialect.
107
108   http://www.w3schools.com/sql/
109      Tutorial, reference and examples for learning SQL syntax.
110
111   :pep:`249` - Database API Specification 2.0
112      PEP written by Marc-André Lemburg.
113
114
115.. _sqlite3-module-contents:
116
117Module functions and constants
118------------------------------
119
120
121.. data:: version
122
123   The version number of this module, as a string. This is not the version of
124   the SQLite library.
125
126.. data:: version_info
127
128   The version number of this module, as a tuple of integers. This is not the
129   version of the SQLite library.
130
131.. data:: sqlite_version
132
133   The version number of the run-time SQLite library, as a string.
134
135.. data:: sqlite_version_info
136
137   The version number of the run-time SQLite library, as a tuple of integers.
138
139.. data:: PARSE_DECLTYPES
140
141   This constant is meant to be used with the *detect_types* parameter of the
142   :func:`connect` function.
143
144   Setting it makes the :mod:`sqlite3` module parse the declared type for each
145   column it returns.  It will parse out the first word of the declared type,
146   i. e.  for "integer primary key", it will parse out "integer", or for
147   "number(10)" it will parse out "number". Then for that column, it will look
148   into the converters dictionary and use the converter function registered for
149   that type there.
150
151
152.. data:: PARSE_COLNAMES
153
154   This constant is meant to be used with the *detect_types* parameter of the
155   :func:`connect` function.
156
157   Setting this makes the SQLite interface parse the column name for each column it
158   returns.  It will look for a string formed [mytype] in there, and then decide
159   that 'mytype' is the type of the column. It will try to find an entry of
160   'mytype' in the converters dictionary and then use the converter function found
161   there to return the value. The column name found in :attr:`Cursor.description`
162   is only the first word of the column name, i.  e. if you use something like
163   ``'as "x [datetime]"'`` in your SQL, then we will parse out everything until the
164   first blank for the column name: the column name would simply be "x".
165
166
167.. function:: connect(database[, timeout, detect_types, isolation_level, check_same_thread, factory, cached_statements])
168
169   Opens a connection to the SQLite database file *database*. You can use
170   ``":memory:"`` to open a database connection to a database that resides in RAM
171   instead of on disk.
172
173   When a database is accessed by multiple connections, and one of the processes
174   modifies the database, the SQLite database is locked until that transaction is
175   committed. The *timeout* parameter specifies how long the connection should wait
176   for the lock to go away until raising an exception. The default for the timeout
177   parameter is 5.0 (five seconds).
178
179   For the *isolation_level* parameter, please see the
180   :attr:`Connection.isolation_level` property of :class:`Connection` objects.
181
182   SQLite natively supports only the types TEXT, INTEGER, REAL, BLOB and NULL. If
183   you want to use other types you must add support for them yourself. The
184   *detect_types* parameter and the using custom **converters** registered with the
185   module-level :func:`register_converter` function allow you to easily do that.
186
187   *detect_types* defaults to 0 (i. e. off, no type detection), you can set it to
188   any combination of :const:`PARSE_DECLTYPES` and :const:`PARSE_COLNAMES` to turn
189   type detection on.
190
191   By default, the :mod:`sqlite3` module uses its :class:`Connection` class for the
192   connect call.  You can, however, subclass the :class:`Connection` class and make
193   :func:`connect` use your class instead by providing your class for the *factory*
194   parameter.
195
196   Consult the section :ref:`sqlite3-types` of this manual for details.
197
198   The :mod:`sqlite3` module internally uses a statement cache to avoid SQL parsing
199   overhead. If you want to explicitly set the number of statements that are cached
200   for the connection, you can set the *cached_statements* parameter. The currently
201   implemented default is to cache 100 statements.
202
203
204.. function:: register_converter(typename, callable)
205
206   Registers a callable to convert a bytestring from the database into a custom
207   Python type. The callable will be invoked for all database values that are of
208   the type *typename*. Confer the parameter *detect_types* of the :func:`connect`
209   function for how the type detection works. Note that the case of *typename* and
210   the name of the type in your query must match!
211
212
213.. function:: register_adapter(type, callable)
214
215   Registers a callable to convert the custom Python type *type* into one of
216   SQLite's supported types. The callable *callable* accepts as single parameter
217   the Python value, and must return a value of the following types: int, long,
218   float, str (UTF-8 encoded), unicode or buffer.
219
220
221.. function:: complete_statement(sql)
222
223   Returns :const:`True` if the string *sql* contains one or more complete SQL
224   statements terminated by semicolons. It does not verify that the SQL is
225   syntactically correct, only that there are no unclosed string literals and the
226   statement is terminated by a semicolon.
227
228   This can be used to build a shell for SQLite, as in the following example:
229
230
231   .. literalinclude:: ../includes/sqlite3/complete_statement.py
232
233
234.. function:: enable_callback_tracebacks(flag)
235
236   By default you will not get any tracebacks in user-defined functions,
237   aggregates, converters, authorizer callbacks etc. If you want to debug them,
238   you can call this function with *flag* set to ``True``. Afterwards, you will
239   get tracebacks from callbacks on ``sys.stderr``. Use :const:`False` to
240   disable the feature again.
241
242
243.. _sqlite3-connection-objects:
244
245Connection Objects
246------------------
247
248.. class:: Connection
249
250   A SQLite database connection has the following attributes and methods:
251
252   .. attribute:: isolation_level
253
254      Get or set the current isolation level. :const:`None` for autocommit mode or
255      one of "DEFERRED", "IMMEDIATE" or "EXCLUSIVE". See section
256      :ref:`sqlite3-controlling-transactions` for a more detailed explanation.
257
258
259   .. method:: cursor(factory=Cursor)
260
261      The cursor method accepts a single optional parameter *factory*. If
262      supplied, this must be a callable returning an instance of :class:`Cursor`
263      or its subclasses.
264
265   .. method:: commit()
266
267      This method commits the current transaction. If you don't call this method,
268      anything you did since the last call to ``commit()`` is not visible from
269      other database connections. If you wonder why you don't see the data you've
270      written to the database, please check you didn't forget to call this method.
271
272   .. method:: rollback()
273
274      This method rolls back any changes to the database since the last call to
275      :meth:`commit`.
276
277   .. method:: close()
278
279      This closes the database connection. Note that this does not automatically
280      call :meth:`commit`. If you just close your database connection without
281      calling :meth:`commit` first, your changes will be lost!
282
283   .. method:: execute(sql, [parameters])
284
285      This is a nonstandard shortcut that creates an intermediate cursor object by
286      calling the cursor method, then calls the cursor's :meth:`execute
287      <Cursor.execute>` method with the parameters given.
288
289
290   .. method:: executemany(sql, [parameters])
291
292      This is a nonstandard shortcut that creates an intermediate cursor object by
293      calling the cursor method, then calls the cursor's :meth:`executemany
294      <Cursor.executemany>` method with the parameters given.
295
296   .. method:: executescript(sql_script)
297
298      This is a nonstandard shortcut that creates an intermediate cursor object by
299      calling the cursor method, then calls the cursor's :meth:`executescript
300      <Cursor.executescript>` method with the parameters given.
301
302
303   .. method:: create_function(name, num_params, func)
304
305      Creates a user-defined function that you can later use from within SQL
306      statements under the function name *name*. *num_params* is the number of
307      parameters the function accepts, and *func* is a Python callable that is called
308      as the SQL function.
309
310      The function can return any of the types supported by SQLite: unicode, str, int,
311      long, float, buffer and ``None``.
312
313      Example:
314
315      .. literalinclude:: ../includes/sqlite3/md5func.py
316
317
318   .. method:: create_aggregate(name, num_params, aggregate_class)
319
320      Creates a user-defined aggregate function.
321
322      The aggregate class must implement a ``step`` method, which accepts the number
323      of parameters *num_params*, and a ``finalize`` method which will return the
324      final result of the aggregate.
325
326      The ``finalize`` method can return any of the types supported by SQLite:
327      unicode, str, int, long, float, buffer and ``None``.
328
329      Example:
330
331      .. literalinclude:: ../includes/sqlite3/mysumaggr.py
332
333
334   .. method:: create_collation(name, callable)
335
336      Creates a collation with the specified *name* and *callable*. The callable will
337      be passed two string arguments. It should return -1 if the first is ordered
338      lower than the second, 0 if they are ordered equal and 1 if the first is ordered
339      higher than the second.  Note that this controls sorting (ORDER BY in SQL) so
340      your comparisons don't affect other SQL operations.
341
342      Note that the callable will get its parameters as Python bytestrings, which will
343      normally be encoded in UTF-8.
344
345      The following example shows a custom collation that sorts "the wrong way":
346
347      .. literalinclude:: ../includes/sqlite3/collation_reverse.py
348
349      To remove a collation, call ``create_collation`` with ``None`` as callable::
350
351         con.create_collation("reverse", None)
352
353
354   .. method:: interrupt()
355
356      You can call this method from a different thread to abort any queries that might
357      be executing on the connection. The query will then abort and the caller will
358      get an exception.
359
360
361   .. method:: set_authorizer(authorizer_callback)
362
363      This routine registers a callback. The callback is invoked for each attempt to
364      access a column of a table in the database. The callback should return
365      :const:`SQLITE_OK` if access is allowed, :const:`SQLITE_DENY` if the entire SQL
366      statement should be aborted with an error and :const:`SQLITE_IGNORE` if the
367      column should be treated as a NULL value. These constants are available in the
368      :mod:`sqlite3` module.
369
370      The first argument to the callback signifies what kind of operation is to be
371      authorized. The second and third argument will be arguments or :const:`None`
372      depending on the first argument. The 4th argument is the name of the database
373      ("main", "temp", etc.) if applicable. The 5th argument is the name of the
374      inner-most trigger or view that is responsible for the access attempt or
375      :const:`None` if this access attempt is directly from input SQL code.
376
377      Please consult the SQLite documentation about the possible values for the first
378      argument and the meaning of the second and third argument depending on the first
379      one. All necessary constants are available in the :mod:`sqlite3` module.
380
381
382   .. method:: set_progress_handler(handler, n)
383
384      This routine registers a callback. The callback is invoked for every *n*
385      instructions of the SQLite virtual machine. This is useful if you want to
386      get called from SQLite during long-running operations, for example to update
387      a GUI.
388
389      If you want to clear any previously installed progress handler, call the
390      method with :const:`None` for *handler*.
391
392      .. versionadded:: 2.6
393
394
395   .. method:: enable_load_extension(enabled)
396
397      This routine allows/disallows the SQLite engine to load SQLite extensions
398      from shared libraries.  SQLite extensions can define new functions,
399      aggregates or whole new virtual table implementations.  One well-known
400      extension is the fulltext-search extension distributed with SQLite.
401
402      Loadable extensions are disabled by default. See [#f1]_.
403
404      .. versionadded:: 2.7
405
406      .. literalinclude:: ../includes/sqlite3/load_extension.py
407
408   .. method:: load_extension(path)
409
410      This routine loads a SQLite extension from a shared library.  You have to
411      enable extension loading with :meth:`enable_load_extension` before you can
412      use this routine.
413
414      Loadable extensions are disabled by default. See [#f1]_.
415
416      .. versionadded:: 2.7
417
418   .. attribute:: row_factory
419
420      You can change this attribute to a callable that accepts the cursor and the
421      original row as a tuple and will return the real result row.  This way, you can
422      implement more advanced ways of returning results, such  as returning an object
423      that can also access columns by name.
424
425      Example:
426
427      .. literalinclude:: ../includes/sqlite3/row_factory.py
428
429      If returning a tuple doesn't suffice and you want name-based access to
430      columns, you should consider setting :attr:`row_factory` to the
431      highly-optimized :class:`sqlite3.Row` type. :class:`Row` provides both
432      index-based and case-insensitive name-based access to columns with almost no
433      memory overhead. It will probably be better than your own custom
434      dictionary-based approach or even a db_row based solution.
435
436      .. XXX what's a db_row-based solution?
437
438
439   .. attribute:: text_factory
440
441      Using this attribute you can control what objects are returned for the ``TEXT``
442      data type. By default, this attribute is set to :class:`unicode` and the
443      :mod:`sqlite3` module will return Unicode objects for ``TEXT``. If you want to
444      return bytestrings instead, you can set it to :class:`str`.
445
446      For efficiency reasons, there's also a way to return Unicode objects only for
447      non-ASCII data, and bytestrings otherwise. To activate it, set this attribute to
448      :const:`sqlite3.OptimizedUnicode`.
449
450      You can also set it to any other callable that accepts a single bytestring
451      parameter and returns the resulting object.
452
453      See the following example code for illustration:
454
455      .. literalinclude:: ../includes/sqlite3/text_factory.py
456
457
458   .. attribute:: total_changes
459
460      Returns the total number of database rows that have been modified, inserted, or
461      deleted since the database connection was opened.
462
463
464   .. attribute:: iterdump
465
466      Returns an iterator to dump the database in an SQL text format.  Useful when
467      saving an in-memory database for later restoration.  This function provides
468      the same capabilities as the :kbd:`.dump` command in the :program:`sqlite3`
469      shell.
470
471      .. versionadded:: 2.6
472
473      Example::
474
475         # Convert file existing_db.db to SQL dump file dump.sql
476         import sqlite3, os
477
478         con = sqlite3.connect('existing_db.db')
479         with open('dump.sql', 'w') as f:
480             for line in con.iterdump():
481                 f.write('%s\n' % line)
482
483
484.. _sqlite3-cursor-objects:
485
486Cursor Objects
487--------------
488
489.. class:: Cursor
490
491   A :class:`Cursor` instance has the following attributes and methods.
492
493   .. method:: execute(sql, [parameters])
494
495      Executes an SQL statement. The SQL statement may be parameterized (i. e.
496      placeholders instead of SQL literals). The :mod:`sqlite3` module supports two
497      kinds of placeholders: question marks (qmark style) and named placeholders
498      (named style).
499
500      Here's an example of both styles:
501
502      .. literalinclude:: ../includes/sqlite3/execute_1.py
503
504      :meth:`execute` will only execute a single SQL statement. If you try to execute
505      more than one statement with it, it will raise a Warning. Use
506      :meth:`executescript` if you want to execute multiple SQL statements with one
507      call.
508
509
510   .. method:: executemany(sql, seq_of_parameters)
511
512      Executes an SQL command against all parameter sequences or mappings found in
513      the sequence *sql*.  The :mod:`sqlite3` module also allows using an
514      :term:`iterator` yielding parameters instead of a sequence.
515
516      .. literalinclude:: ../includes/sqlite3/executemany_1.py
517
518      Here's a shorter example using a :term:`generator`:
519
520      .. literalinclude:: ../includes/sqlite3/executemany_2.py
521
522
523   .. method:: executescript(sql_script)
524
525      This is a nonstandard convenience method for executing multiple SQL statements
526      at once. It issues a ``COMMIT`` statement first, then executes the SQL script it
527      gets as a parameter.
528
529      *sql_script* can be a bytestring or a Unicode string.
530
531      Example:
532
533      .. literalinclude:: ../includes/sqlite3/executescript.py
534
535
536   .. method:: fetchone()
537
538      Fetches the next row of a query result set, returning a single sequence,
539      or :const:`None` when no more data is available.
540
541
542   .. method:: fetchmany([size=cursor.arraysize])
543
544      Fetches the next set of rows of a query result, returning a list.  An empty
545      list is returned when no more rows are available.
546
547      The number of rows to fetch per call is specified by the *size* parameter.
548      If it is not given, the cursor's arraysize determines the number of rows
549      to be fetched. The method should try to fetch as many rows as indicated by
550      the size parameter. If this is not possible due to the specified number of
551      rows not being available, fewer rows may be returned.
552
553      Note there are performance considerations involved with the *size* parameter.
554      For optimal performance, it is usually best to use the arraysize attribute.
555      If the *size* parameter is used, then it is best for it to retain the same
556      value from one :meth:`fetchmany` call to the next.
557
558   .. method:: fetchall()
559
560      Fetches all (remaining) rows of a query result, returning a list.  Note that
561      the cursor's arraysize attribute can affect the performance of this operation.
562      An empty list is returned when no rows are available.
563
564
565   .. attribute:: rowcount
566
567      Although the :class:`Cursor` class of the :mod:`sqlite3` module implements this
568      attribute, the database engine's own support for the determination of "rows
569      affected"/"rows selected" is quirky.
570
571      For :meth:`executemany` statements, the number of modifications are summed up
572      into :attr:`rowcount`.
573
574      As required by the Python DB API Spec, the :attr:`rowcount` attribute "is -1 in
575      case no ``executeXX()`` has been performed on the cursor or the rowcount of the
576      last operation is not determinable by the interface". This includes ``SELECT``
577      statements because we cannot determine the number of rows a query produced
578      until all rows were fetched.
579
580      With SQLite versions before 3.6.5, :attr:`rowcount` is set to 0 if
581      you make a ``DELETE FROM table`` without any condition.
582
583   .. attribute:: lastrowid
584
585      This read-only attribute provides the rowid of the last modified row. It is
586      only set if you issued an ``INSERT`` statement using the :meth:`execute`
587      method. For operations other than ``INSERT`` or when :meth:`executemany` is
588      called, :attr:`lastrowid` is set to :const:`None`.
589
590   .. attribute:: description
591
592      This read-only attribute provides the column names of the last query. To
593      remain compatible with the Python DB API, it returns a 7-tuple for each
594      column where the last six items of each tuple are :const:`None`.
595
596      It is set for ``SELECT`` statements without any matching rows as well.
597
598   .. attribute:: connection
599
600      This read-only attribute provides the SQLite database :class:`Connection`
601      used by the :class:`Cursor` object.  A :class:`Cursor` object created by
602      calling :meth:`con.cursor() <Connection.cursor>` will have a
603      :attr:`connection` attribute that refers to *con*::
604
605         >>> con = sqlite3.connect(":memory:")
606         >>> cur = con.cursor()
607         >>> cur.connection == con
608         True
609
610.. _sqlite3-row-objects:
611
612Row Objects
613-----------
614
615.. class:: Row
616
617   A :class:`Row` instance serves as a highly optimized
618   :attr:`~Connection.row_factory` for :class:`Connection` objects.
619   It tries to mimic a tuple in most of its features.
620
621   It supports mapping access by column name and index, iteration,
622   representation, equality testing and :func:`len`.
623
624   If two :class:`Row` objects have exactly the same columns and their
625   members are equal, they compare equal.
626
627   .. versionchanged:: 2.6
628      Added iteration and equality (hashability).
629
630   .. method:: keys
631
632      This method returns a list of column names. Immediately after a query,
633      it is the first member of each tuple in :attr:`Cursor.description`.
634
635      .. versionadded:: 2.6
636
637Let's assume we initialize a table as in the example given above::
638
639   conn = sqlite3.connect(":memory:")
640   c = conn.cursor()
641   c.execute('''create table stocks
642   (date text, trans text, symbol text,
643    qty real, price real)''')
644   c.execute("""insert into stocks
645             values ('2006-01-05','BUY','RHAT',100,35.14)""")
646   conn.commit()
647   c.close()
648
649Now we plug :class:`Row` in::
650
651   >>> conn.row_factory = sqlite3.Row
652   >>> c = conn.cursor()
653   >>> c.execute('select * from stocks')
654   <sqlite3.Cursor object at 0x7f4e7dd8fa80>
655   >>> r = c.fetchone()
656   >>> type(r)
657   <type 'sqlite3.Row'>
658   >>> r
659   (u'2006-01-05', u'BUY', u'RHAT', 100.0, 35.14)
660   >>> len(r)
661   5
662   >>> r[2]
663   u'RHAT'
664   >>> r.keys()
665   ['date', 'trans', 'symbol', 'qty', 'price']
666   >>> r['qty']
667   100.0
668   >>> for member in r:
669   ...     print member
670   ...
671   2006-01-05
672   BUY
673   RHAT
674   100.0
675   35.14
676
677
678.. _sqlite3-types:
679
680SQLite and Python types
681-----------------------
682
683
684Introduction
685^^^^^^^^^^^^
686
687SQLite natively supports the following types: ``NULL``, ``INTEGER``,
688``REAL``, ``TEXT``, ``BLOB``.
689
690The following Python types can thus be sent to SQLite without any problem:
691
692+-----------------------------+-------------+
693| Python type                 | SQLite type |
694+=============================+=============+
695| :const:`None`               | ``NULL``    |
696+-----------------------------+-------------+
697| :class:`int`                | ``INTEGER`` |
698+-----------------------------+-------------+
699| :class:`long`               | ``INTEGER`` |
700+-----------------------------+-------------+
701| :class:`float`              | ``REAL``    |
702+-----------------------------+-------------+
703| :class:`str` (UTF8-encoded) | ``TEXT``    |
704+-----------------------------+-------------+
705| :class:`unicode`            | ``TEXT``    |
706+-----------------------------+-------------+
707| :class:`buffer`             | ``BLOB``    |
708+-----------------------------+-------------+
709
710This is how SQLite types are converted to Python types by default:
711
712+-------------+----------------------------------------------+
713| SQLite type | Python type                                  |
714+=============+==============================================+
715| ``NULL``    | :const:`None`                                |
716+-------------+----------------------------------------------+
717| ``INTEGER`` | :class:`int` or :class:`long`,               |
718|             | depending on size                            |
719+-------------+----------------------------------------------+
720| ``REAL``    | :class:`float`                               |
721+-------------+----------------------------------------------+
722| ``TEXT``    | depends on :attr:`~Connection.text_factory`, |
723|             | :class:`unicode` by default                  |
724+-------------+----------------------------------------------+
725| ``BLOB``    | :class:`buffer`                              |
726+-------------+----------------------------------------------+
727
728The type system of the :mod:`sqlite3` module is extensible in two ways: you can
729store additional Python types in a SQLite database via object adaptation, and
730you can let the :mod:`sqlite3` module convert SQLite types to different Python
731types via converters.
732
733
734Using adapters to store additional Python types in SQLite databases
735^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
736
737As described before, SQLite supports only a limited set of types natively. To
738use other Python types with SQLite, you must **adapt** them to one of the
739sqlite3 module's supported types for SQLite: one of NoneType, int, long, float,
740str, unicode, buffer.
741
742There are two ways to enable the :mod:`sqlite3` module to adapt a custom Python
743type to one of the supported ones.
744
745
746Letting your object adapt itself
747""""""""""""""""""""""""""""""""
748
749This is a good approach if you write the class yourself. Let's suppose you have
750a class like this::
751
752   class Point(object):
753       def __init__(self, x, y):
754           self.x, self.y = x, y
755
756Now you want to store the point in a single SQLite column.  First you'll have to
757choose one of the supported types first to be used for representing the point.
758Let's just use str and separate the coordinates using a semicolon. Then you need
759to give your class a method ``__conform__(self, protocol)`` which must return
760the converted value. The parameter *protocol* will be :class:`PrepareProtocol`.
761
762.. literalinclude:: ../includes/sqlite3/adapter_point_1.py
763
764
765Registering an adapter callable
766"""""""""""""""""""""""""""""""
767
768The other possibility is to create a function that converts the type to the
769string representation and register the function with :meth:`register_adapter`.
770
771.. note::
772
773   The type/class to adapt must be a :term:`new-style class`, i. e. it must have
774   :class:`object` as one of its bases.
775
776.. literalinclude:: ../includes/sqlite3/adapter_point_2.py
777
778The :mod:`sqlite3` module has two default adapters for Python's built-in
779:class:`datetime.date` and :class:`datetime.datetime` types.  Now let's suppose
780we want to store :class:`datetime.datetime` objects not in ISO representation,
781but as a Unix timestamp.
782
783.. literalinclude:: ../includes/sqlite3/adapter_datetime.py
784
785
786Converting SQLite values to custom Python types
787^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
788
789Writing an adapter lets you send custom Python types to SQLite. But to make it
790really useful we need to make the Python to SQLite to Python roundtrip work.
791
792Enter converters.
793
794Let's go back to the :class:`Point` class. We stored the x and y coordinates
795separated via semicolons as strings in SQLite.
796
797First, we'll define a converter function that accepts the string as a parameter
798and constructs a :class:`Point` object from it.
799
800.. note::
801
802   Converter functions **always** get called with a string, no matter under which
803   data type you sent the value to SQLite.
804
805::
806
807   def convert_point(s):
808       x, y = map(float, s.split(";"))
809       return Point(x, y)
810
811Now you need to make the :mod:`sqlite3` module know that what you select from
812the database is actually a point. There are two ways of doing this:
813
814* Implicitly via the declared type
815
816* Explicitly via the column name
817
818Both ways are described in section :ref:`sqlite3-module-contents`, in the entries
819for the constants :const:`PARSE_DECLTYPES` and :const:`PARSE_COLNAMES`.
820
821The following example illustrates both approaches.
822
823.. literalinclude:: ../includes/sqlite3/converter_point.py
824
825
826Default adapters and converters
827^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
828
829There are default adapters for the date and datetime types in the datetime
830module. They will be sent as ISO dates/ISO timestamps to SQLite.
831
832The default converters are registered under the name "date" for
833:class:`datetime.date` and under the name "timestamp" for
834:class:`datetime.datetime`.
835
836This way, you can use date/timestamps from Python without any additional
837fiddling in most cases. The format of the adapters is also compatible with the
838experimental SQLite date/time functions.
839
840The following example demonstrates this.
841
842.. literalinclude:: ../includes/sqlite3/pysqlite_datetime.py
843
844If a timestamp stored in SQLite has a fractional part longer than 6
845numbers, its value will be truncated to microsecond precision by the
846timestamp converter.
847
848
849.. _sqlite3-controlling-transactions:
850
851Controlling Transactions
852------------------------
853
854By default, the :mod:`sqlite3` module opens transactions implicitly before a
855Data Modification Language (DML)  statement (i.e.
856``INSERT``/``UPDATE``/``DELETE``/``REPLACE``), and commits transactions
857implicitly before a non-DML, non-query statement (i. e.
858anything other than ``SELECT`` or the aforementioned).
859
860So if you are within a transaction and issue a command like ``CREATE TABLE
861...``, ``VACUUM``, ``PRAGMA``, the :mod:`sqlite3` module will commit implicitly
862before executing that command. There are two reasons for doing that. The first
863is that some of these commands don't work within transactions. The other reason
864is that sqlite3 needs to keep track of the transaction state (if a transaction
865is active or not).
866
867You can control which kind of ``BEGIN`` statements sqlite3 implicitly executes
868(or none at all) via the *isolation_level* parameter to the :func:`connect`
869call, or via the :attr:`isolation_level` property of connections.
870
871If you want **autocommit mode**, then set :attr:`isolation_level` to ``None``.
872
873Otherwise leave it at its default, which will result in a plain "BEGIN"
874statement, or set it to one of SQLite's supported isolation levels: "DEFERRED",
875"IMMEDIATE" or "EXCLUSIVE".
876
877
878
879Using :mod:`sqlite3` efficiently
880--------------------------------
881
882
883Using shortcut methods
884^^^^^^^^^^^^^^^^^^^^^^
885
886Using the nonstandard :meth:`execute`, :meth:`executemany` and
887:meth:`executescript` methods of the :class:`Connection` object, your code can
888be written more concisely because you don't have to create the (often
889superfluous) :class:`Cursor` objects explicitly. Instead, the :class:`Cursor`
890objects are created implicitly and these shortcut methods return the cursor
891objects. This way, you can execute a ``SELECT`` statement and iterate over it
892directly using only a single call on the :class:`Connection` object.
893
894.. literalinclude:: ../includes/sqlite3/shortcut_methods.py
895
896
897Accessing columns by name instead of by index
898^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
899
900One useful feature of the :mod:`sqlite3` module is the built-in
901:class:`sqlite3.Row` class designed to be used as a row factory.
902
903Rows wrapped with this class can be accessed both by index (like tuples) and
904case-insensitively by name:
905
906.. literalinclude:: ../includes/sqlite3/rowclass.py
907
908
909Using the connection as a context manager
910^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
911
912.. versionadded:: 2.6
913
914Connection objects can be used as context managers
915that automatically commit or rollback transactions.  In the event of an
916exception, the transaction is rolled back; otherwise, the transaction is
917committed:
918
919.. literalinclude:: ../includes/sqlite3/ctx_manager.py
920
921
922Common issues
923-------------
924
925Multithreading
926^^^^^^^^^^^^^^
927
928Older SQLite versions had issues with sharing connections between threads.
929That's why the Python module disallows sharing connections and cursors between
930threads. If you still try to do so, you will get an exception at runtime.
931
932The only exception is calling the :meth:`~Connection.interrupt` method, which
933only makes sense to call from a different thread.
934
935.. rubric:: Footnotes
936
937.. [#f1] The sqlite3 module is not built with loadable extension support by
938   default, because some platforms (notably Mac OS X) have SQLite libraries
939   which are compiled without this feature. To get loadable extension support,
940   you must modify setup.py and remove the line that sets
941   SQLITE_OMIT_LOAD_EXTENSION.
942
943