blob: dbef60f0a6b69701719991a65366798b39b05545 [file] [log] [blame]
Georg Brandl8ec7f652007-08-15 14:28:01 +00001
2:mod:`csv` --- CSV File Reading and Writing
3===========================================
4
5.. module:: csv
6 :synopsis: Write and read tabular data to and from delimited files.
7.. sectionauthor:: Skip Montanaro <skip@pobox.com>
8
9
10.. versionadded:: 2.3
11
12.. index::
13 single: csv
14 pair: data; tabular
15
16The so-called CSV (Comma Separated Values) format is the most common import and
17export format for spreadsheets and databases. There is no "CSV standard", so
18the format is operationally defined by the many applications which read and
19write it. The lack of a standard means that subtle differences often exist in
20the data produced and consumed by different applications. These differences can
21make it annoying to process CSV files from multiple sources. Still, while the
22delimiters and quoting characters vary, the overall format is similar enough
23that it is possible to write a single module which can efficiently manipulate
24such data, hiding the details of reading and writing the data from the
25programmer.
26
27The :mod:`csv` module implements classes to read and write tabular data in CSV
28format. It allows programmers to say, "write this data in the format preferred
29by Excel," or "read data from this file which was generated by Excel," without
30knowing the precise details of the CSV format used by Excel. Programmers can
31also describe the CSV formats understood by other applications or define their
32own special-purpose CSV formats.
33
34The :mod:`csv` module's :class:`reader` and :class:`writer` objects read and
35write sequences. Programmers can also read and write data in dictionary form
36using the :class:`DictReader` and :class:`DictWriter` classes.
37
38.. note::
39
40 This version of the :mod:`csv` module doesn't support Unicode input. Also,
41 there are currently some issues regarding ASCII NUL characters. Accordingly,
42 all input should be UTF-8 or printable ASCII to be safe; see the examples in
Éric Araujo06176a82012-07-02 17:46:40 -040043 section :ref:`csv-examples`.
Georg Brandl8ec7f652007-08-15 14:28:01 +000044
45
46.. seealso::
47
Georg Brandl8ec7f652007-08-15 14:28:01 +000048 :pep:`305` - CSV File API
49 The Python Enhancement Proposal which proposed this addition to Python.
50
51
52.. _csv-contents:
53
54Module Contents
55---------------
56
57The :mod:`csv` module defines the following functions:
58
59
Hynek Schlawack7d978902012-08-28 12:33:46 +020060.. function:: reader(csvfile, dialect='excel', **fmtparams)
Georg Brandl8ec7f652007-08-15 14:28:01 +000061
62 Return a reader object which will iterate over lines in the given *csvfile*.
Georg Brandle7a09902007-10-21 12:10:28 +000063 *csvfile* can be any object which supports the :term:`iterator` protocol and returns a
Georg Brandl9fa61bb2009-07-26 14:19:57 +000064 string each time its :meth:`!next` method is called --- file objects and list
Georg Brandl8ec7f652007-08-15 14:28:01 +000065 objects are both suitable. If *csvfile* is a file object, it must be opened
66 with the 'b' flag on platforms where that makes a difference. An optional
67 *dialect* parameter can be given which is used to define a set of parameters
68 specific to a particular CSV dialect. It may be an instance of a subclass of
69 the :class:`Dialect` class or one of the strings returned by the
Hynek Schlawack7d978902012-08-28 12:33:46 +020070 :func:`list_dialects` function. The other optional *fmtparams* keyword arguments
Georg Brandl8ec7f652007-08-15 14:28:01 +000071 can be given to override individual formatting parameters in the current
72 dialect. For full details about the dialect and formatting parameters, see
73 section :ref:`csv-fmt-params`.
74
Skip Montanaro9a1337b2009-03-25 00:52:11 +000075 Each row read from the csv file is returned as a list of strings. No
76 automatic data type conversion is performed.
Georg Brandl8ec7f652007-08-15 14:28:01 +000077
Georg Brandl722e1012007-12-05 17:56:50 +000078 A short usage example::
Georg Brandlc62ef8b2009-01-03 20:55:06 +000079
Georg Brandl722e1012007-12-05 17:56:50 +000080 >>> import csv
Ezio Melottia733d812012-09-15 05:46:24 +030081 >>> with open('eggs.csv', 'rb') as csvfile:
82 ... spamreader = csv.reader(csvfile, delimiter=' ', quotechar='|')
83 ... for row in spamreader:
84 ... print ', '.join(row)
Georg Brandl722e1012007-12-05 17:56:50 +000085 Spam, Spam, Spam, Spam, Spam, Baked Beans
86 Spam, Lovely Spam, Wonderful Spam
87
Georg Brandl8ec7f652007-08-15 14:28:01 +000088 .. versionchanged:: 2.5
89 The parser is now stricter with respect to multi-line quoted fields. Previously,
90 if a line ended within a quoted field without a terminating newline character, a
91 newline would be inserted into the returned field. This behavior caused problems
92 when reading files which contained carriage return characters within fields.
93 The behavior was changed to return the field without inserting newlines. As a
94 consequence, if newlines embedded within fields are important, the input should
95 be split into lines in a manner which preserves the newline characters.
96
97
Hynek Schlawack7d978902012-08-28 12:33:46 +020098.. function:: writer(csvfile, dialect='excel', **fmtparams)
Georg Brandl8ec7f652007-08-15 14:28:01 +000099
100 Return a writer object responsible for converting the user's data into delimited
101 strings on the given file-like object. *csvfile* can be any object with a
102 :func:`write` method. If *csvfile* is a file object, it must be opened with the
103 'b' flag on platforms where that makes a difference. An optional *dialect*
104 parameter can be given which is used to define a set of parameters specific to a
105 particular CSV dialect. It may be an instance of a subclass of the
106 :class:`Dialect` class or one of the strings returned by the
Hynek Schlawack7d978902012-08-28 12:33:46 +0200107 :func:`list_dialects` function. The other optional *fmtparams* keyword arguments
Georg Brandl8ec7f652007-08-15 14:28:01 +0000108 can be given to override individual formatting parameters in the current
109 dialect. For full details about the dialect and formatting parameters, see
110 section :ref:`csv-fmt-params`. To make it
111 as easy as possible to interface with modules which implement the DB API, the
112 value :const:`None` is written as the empty string. While this isn't a
113 reversible transformation, it makes it easier to dump SQL NULL data values to
114 CSV files without preprocessing the data returned from a ``cursor.fetch*`` call.
Raymond Hettinger5a3e8e52016-02-27 23:34:54 -0800115 Floats are stringified with :func:`repr` before being written.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000116 All other non-string data are stringified with :func:`str` before being written.
117
Georg Brandl722e1012007-12-05 17:56:50 +0000118 A short usage example::
119
Ezio Melottia733d812012-09-15 05:46:24 +0300120 import csv
121 with open('eggs.csv', 'wb') as csvfile:
122 spamwriter = csv.writer(csvfile, delimiter=' ',
123 quotechar='|', quoting=csv.QUOTE_MINIMAL)
124 spamwriter.writerow(['Spam'] * 5 + ['Baked Beans'])
125 spamwriter.writerow(['Spam', 'Lovely Spam', 'Wonderful Spam'])
Georg Brandl722e1012007-12-05 17:56:50 +0000126
Georg Brandl8ec7f652007-08-15 14:28:01 +0000127
Hynek Schlawack7d978902012-08-28 12:33:46 +0200128.. function:: register_dialect(name[, dialect], **fmtparams)
Georg Brandl8ec7f652007-08-15 14:28:01 +0000129
130 Associate *dialect* with *name*. *name* must be a string or Unicode object. The
131 dialect can be specified either by passing a sub-class of :class:`Dialect`, or
Hynek Schlawack7d978902012-08-28 12:33:46 +0200132 by *fmtparams* keyword arguments, or both, with keyword arguments overriding
Georg Brandl8ec7f652007-08-15 14:28:01 +0000133 parameters of the dialect. For full details about the dialect and formatting
134 parameters, see section :ref:`csv-fmt-params`.
135
136
137.. function:: unregister_dialect(name)
138
139 Delete the dialect associated with *name* from the dialect registry. An
140 :exc:`Error` is raised if *name* is not a registered dialect name.
141
142
143.. function:: get_dialect(name)
144
145 Return the dialect associated with *name*. An :exc:`Error` is raised if *name*
146 is not a registered dialect name.
147
Skip Montanarod469ff12007-11-04 15:56:52 +0000148 .. versionchanged:: 2.5
Georg Brandl9c466ba2007-11-04 17:43:49 +0000149 This function now returns an immutable :class:`Dialect`. Previously an
150 instance of the requested dialect was returned. Users could modify the
151 underlying class, changing the behavior of active readers and writers.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000152
153.. function:: list_dialects()
154
155 Return the names of all registered dialects.
156
157
158.. function:: field_size_limit([new_limit])
159
160 Returns the current maximum field size allowed by the parser. If *new_limit* is
161 given, this becomes the new limit.
162
163 .. versionadded:: 2.5
164
165The :mod:`csv` module defines the following classes:
166
167
R David Murray17a43222014-02-24 15:36:45 -0500168.. class:: DictReader(csvfile, fieldnames=None, restkey=None, restval=None, \
169 dialect='excel', *args, **kwds)
Georg Brandl8ec7f652007-08-15 14:28:01 +0000170
R David Murray17a43222014-02-24 15:36:45 -0500171 Create an object which operates like a regular reader but maps the
172 information read into a dict whose keys are given by the optional
173 *fieldnames* parameter. The *fieldnames* parameter is a :ref:`sequence
R David Murrayd2b5b312014-02-24 15:35:19 -0500174 <collections-abstract-base-classes>` whose elements are associated with the
175 fields of the input data in order. These elements become the keys of the
R David Murray17a43222014-02-24 15:36:45 -0500176 resulting dictionary. If the *fieldnames* parameter is omitted, the values
177 in the first row of the *csvfile* will be used as the fieldnames. If the
178 row read has more fields than the fieldnames sequence, the remaining data is
179 added as a sequence keyed by the value of *restkey*. If the row read has
180 fewer fields than the fieldnames sequence, the remaining keys take the value
181 of the optional *restval* parameter. Any other optional or keyword
182 arguments are passed to the underlying :class:`reader` instance.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000183
Berker Peksag8a9c6822014-11-24 23:50:46 +0200184 A short usage example::
185
186 >>> import csv
187 >>> with open('names.csv') as csvfile:
188 ... reader = csv.DictReader(csvfile)
189 ... for row in reader:
190 ... print(row['first_name'], row['last_name'])
191 ...
192 Baked Beans
193 Lovely Spam
194 Wonderful Spam
195
Georg Brandl8ec7f652007-08-15 14:28:01 +0000196
R David Murray17a43222014-02-24 15:36:45 -0500197.. class:: DictWriter(csvfile, fieldnames, restval='', extrasaction='raise', \
198 dialect='excel', *args, **kwds)
Georg Brandl8ec7f652007-08-15 14:28:01 +0000199
R David Murray17a43222014-02-24 15:36:45 -0500200 Create an object which operates like a regular writer but maps dictionaries
201 onto output rows. The *fieldnames* parameter is a :ref:`sequence
R David Murrayd2b5b312014-02-24 15:35:19 -0500202 <collections-abstract-base-classes>` of keys that identify the order in
R David Murray17a43222014-02-24 15:36:45 -0500203 which values in the dictionary passed to the :meth:`writerow` method are
204 written to the *csvfile*. The optional *restval* parameter specifies the
205 value to be written if the dictionary is missing a key in *fieldnames*. If
206 the dictionary passed to the :meth:`writerow` method contains a key not
207 found in *fieldnames*, the optional *extrasaction* parameter indicates what
208 action to take. If it is set to ``'raise'`` a :exc:`ValueError` is raised.
209 If it is set to ``'ignore'``, extra values in the dictionary are ignored.
210 Any other optional or keyword arguments are passed to the underlying
211 :class:`writer` instance.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000212
R David Murray17a43222014-02-24 15:36:45 -0500213 Note that unlike the :class:`DictReader` class, the *fieldnames* parameter
214 of the :class:`DictWriter` is not optional. Since Python's :class:`dict`
215 objects are not ordered, there is not enough information available to deduce
216 the order in which the row should be written to the *csvfile*.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000217
Berker Peksag8a9c6822014-11-24 23:50:46 +0200218 A short usage example::
219
220 import csv
221
222 with open('names.csv', 'w') as csvfile:
223 fieldnames = ['first_name', 'last_name']
224 writer = csv.DictWriter(csvfile, fieldnames=fieldnames)
225
226 writer.writeheader()
227 writer.writerow({'first_name': 'Baked', 'last_name': 'Beans'})
228 writer.writerow({'first_name': 'Lovely', 'last_name': 'Spam'})
229 writer.writerow({'first_name': 'Wonderful', 'last_name': 'Spam'})
230
Georg Brandl8ec7f652007-08-15 14:28:01 +0000231
232.. class:: Dialect
233
234 The :class:`Dialect` class is a container class relied on primarily for its
235 attributes, which are used to define the parameters for a specific
236 :class:`reader` or :class:`writer` instance.
237
238
239.. class:: excel()
240
241 The :class:`excel` class defines the usual properties of an Excel-generated CSV
242 file. It is registered with the dialect name ``'excel'``.
243
244
245.. class:: excel_tab()
246
247 The :class:`excel_tab` class defines the usual properties of an Excel-generated
248 TAB-delimited file. It is registered with the dialect name ``'excel-tab'``.
249
250
251.. class:: Sniffer()
252
253 The :class:`Sniffer` class is used to deduce the format of a CSV file.
254
Benjamin Petersonc7b05922008-04-25 01:29:10 +0000255 The :class:`Sniffer` class provides two methods:
Georg Brandl8ec7f652007-08-15 14:28:01 +0000256
Hynek Schlawacke58ce012012-05-22 10:27:40 +0200257 .. method:: sniff(sample, delimiters=None)
Georg Brandl8ec7f652007-08-15 14:28:01 +0000258
Benjamin Petersonc7b05922008-04-25 01:29:10 +0000259 Analyze the given *sample* and return a :class:`Dialect` subclass
260 reflecting the parameters found. If the optional *delimiters* parameter
261 is given, it is interpreted as a string containing possible valid
262 delimiter characters.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000263
264
Benjamin Petersonc7b05922008-04-25 01:29:10 +0000265 .. method:: has_header(sample)
Georg Brandl8ec7f652007-08-15 14:28:01 +0000266
Benjamin Petersonc7b05922008-04-25 01:29:10 +0000267 Analyze the sample text (presumed to be in CSV format) and return
268 :const:`True` if the first row appears to be a series of column headers.
Georg Brandl8ec7f652007-08-15 14:28:01 +0000269
Georg Brandl14aaee12008-01-06 16:04:56 +0000270An example for :class:`Sniffer` use::
Georg Brandl8ec7f652007-08-15 14:28:01 +0000271
Ezio Melottia733d812012-09-15 05:46:24 +0300272 with open('example.csv', 'rb') as csvfile:
273 dialect = csv.Sniffer().sniff(csvfile.read(1024))
274 csvfile.seek(0)
275 reader = csv.reader(csvfile, dialect)
276 # ... process CSV file contents here ...
Georg Brandl14aaee12008-01-06 16:04:56 +0000277
278
279The :mod:`csv` module defines the following constants:
Georg Brandl8ec7f652007-08-15 14:28:01 +0000280
281.. data:: QUOTE_ALL
282
283 Instructs :class:`writer` objects to quote all fields.
284
285
286.. data:: QUOTE_MINIMAL
287
288 Instructs :class:`writer` objects to only quote those fields which contain
289 special characters such as *delimiter*, *quotechar* or any of the characters in
290 *lineterminator*.
291
292
293.. data:: QUOTE_NONNUMERIC
294
295 Instructs :class:`writer` objects to quote all non-numeric fields.
296
297 Instructs the reader to convert all non-quoted fields to type *float*.
298
299
300.. data:: QUOTE_NONE
301
302 Instructs :class:`writer` objects to never quote fields. When the current
303 *delimiter* occurs in output data it is preceded by the current *escapechar*
304 character. If *escapechar* is not set, the writer will raise :exc:`Error` if
305 any characters that require escaping are encountered.
306
307 Instructs :class:`reader` to perform no special processing of quote characters.
308
309The :mod:`csv` module defines the following exception:
310
311
312.. exception:: Error
313
314 Raised by any of the functions when an error is detected.
315
316
317.. _csv-fmt-params:
318
319Dialects and Formatting Parameters
320----------------------------------
321
322To make it easier to specify the format of input and output records, specific
323formatting parameters are grouped together into dialects. A dialect is a
324subclass of the :class:`Dialect` class having a set of specific methods and a
325single :meth:`validate` method. When creating :class:`reader` or
326:class:`writer` objects, the programmer can specify a string or a subclass of
327the :class:`Dialect` class as the dialect parameter. In addition to, or instead
328of, the *dialect* parameter, the programmer can also specify individual
329formatting parameters, which have the same names as the attributes defined below
330for the :class:`Dialect` class.
331
332Dialects support the following attributes:
333
334
335.. attribute:: Dialect.delimiter
336
337 A one-character string used to separate fields. It defaults to ``','``.
338
339
340.. attribute:: Dialect.doublequote
341
Zachary Wared0384042015-09-11 10:51:47 -0500342 Controls how instances of *quotechar* appearing inside a field should
Georg Brandl8ec7f652007-08-15 14:28:01 +0000343 themselves be quoted. When :const:`True`, the character is doubled. When
344 :const:`False`, the *escapechar* is used as a prefix to the *quotechar*. It
345 defaults to :const:`True`.
346
347 On output, if *doublequote* is :const:`False` and no *escapechar* is set,
348 :exc:`Error` is raised if a *quotechar* is found in a field.
349
350
351.. attribute:: Dialect.escapechar
352
353 A one-character string used by the writer to escape the *delimiter* if *quoting*
354 is set to :const:`QUOTE_NONE` and the *quotechar* if *doublequote* is
355 :const:`False`. On reading, the *escapechar* removes any special meaning from
356 the following character. It defaults to :const:`None`, which disables escaping.
357
358
359.. attribute:: Dialect.lineterminator
360
361 The string used to terminate lines produced by the :class:`writer`. It defaults
362 to ``'\r\n'``.
363
364 .. note::
365
366 The :class:`reader` is hard-coded to recognise either ``'\r'`` or ``'\n'`` as
367 end-of-line, and ignores *lineterminator*. This behavior may change in the
368 future.
369
370
371.. attribute:: Dialect.quotechar
372
373 A one-character string used to quote fields containing special characters, such
374 as the *delimiter* or *quotechar*, or which contain new-line characters. It
375 defaults to ``'"'``.
376
377
378.. attribute:: Dialect.quoting
379
380 Controls when quotes should be generated by the writer and recognised by the
381 reader. It can take on any of the :const:`QUOTE_\*` constants (see section
382 :ref:`csv-contents`) and defaults to :const:`QUOTE_MINIMAL`.
383
384
385.. attribute:: Dialect.skipinitialspace
386
387 When :const:`True`, whitespace immediately following the *delimiter* is ignored.
388 The default is :const:`False`.
389
390
Ezio Melotti355637b2012-11-18 12:55:35 +0200391.. attribute:: Dialect.strict
392
393 When ``True``, raise exception :exc:`Error` on bad CSV input.
394 The default is ``False``.
395
Georg Brandl8ec7f652007-08-15 14:28:01 +0000396Reader Objects
397--------------
398
399Reader objects (:class:`DictReader` instances and objects returned by the
400:func:`reader` function) have the following public methods:
401
402
403.. method:: csvreader.next()
404
405 Return the next row of the reader's iterable object as a list, parsed according
406 to the current dialect.
407
408Reader objects have the following public attributes:
409
410
411.. attribute:: csvreader.dialect
412
413 A read-only description of the dialect in use by the parser.
414
415
416.. attribute:: csvreader.line_num
417
418 The number of lines read from the source iterator. This is not the same as the
419 number of records returned, as records can span multiple lines.
420
421 .. versionadded:: 2.5
422
423
Skip Montanaroa032bf42008-08-08 22:52:51 +0000424DictReader objects have the following public attribute:
425
426
427.. attribute:: csvreader.fieldnames
428
429 If not passed as a parameter when creating the object, this attribute is
430 initialized upon first access or when the first record is read from the
431 file.
432
433 .. versionchanged:: 2.6
434
435
Georg Brandl8ec7f652007-08-15 14:28:01 +0000436Writer Objects
437--------------
438
439:class:`Writer` objects (:class:`DictWriter` instances and objects returned by
440the :func:`writer` function) have the following public methods. A *row* must be
441a sequence of strings or numbers for :class:`Writer` objects and a dictionary
442mapping fieldnames to strings or numbers (by passing them through :func:`str`
443first) for :class:`DictWriter` objects. Note that complex numbers are written
444out surrounded by parens. This may cause some problems for other programs which
445read CSV files (assuming they support complex numbers at all).
446
447
448.. method:: csvwriter.writerow(row)
449
450 Write the *row* parameter to the writer's file object, formatted according to
451 the current dialect.
452
453
454.. method:: csvwriter.writerows(rows)
455
456 Write all the *rows* parameters (a list of *row* objects as described above) to
457 the writer's file object, formatted according to the current dialect.
458
459Writer objects have the following public attribute:
460
461
462.. attribute:: csvwriter.dialect
463
464 A read-only description of the dialect in use by the writer.
465
466
Dirkjan Ochtman86148172010-02-23 21:09:52 +0000467DictWriter objects have the following public method:
468
469
470.. method:: DictWriter.writeheader()
471
472 Write a row with the field names (as specified in the constructor).
473
474 .. versionadded:: 2.7
475
476
Georg Brandl8ec7f652007-08-15 14:28:01 +0000477.. _csv-examples:
478
479Examples
480--------
481
482The simplest example of reading a CSV file::
483
484 import csv
Eli Benderskyec40bab2011-03-13 08:45:19 +0200485 with open('some.csv', 'rb') as f:
486 reader = csv.reader(f)
487 for row in reader:
488 print row
Georg Brandl8ec7f652007-08-15 14:28:01 +0000489
490Reading a file with an alternate format::
491
492 import csv
Eli Benderskyec40bab2011-03-13 08:45:19 +0200493 with open('passwd', 'rb') as f:
494 reader = csv.reader(f, delimiter=':', quoting=csv.QUOTE_NONE)
495 for row in reader:
496 print row
Georg Brandl8ec7f652007-08-15 14:28:01 +0000497
498The corresponding simplest possible writing example is::
499
500 import csv
Eli Benderskyec40bab2011-03-13 08:45:19 +0200501 with open('some.csv', 'wb') as f:
502 writer = csv.writer(f)
503 writer.writerows(someiterable)
Georg Brandl8ec7f652007-08-15 14:28:01 +0000504
505Registering a new dialect::
506
507 import csv
Georg Brandl8ec7f652007-08-15 14:28:01 +0000508 csv.register_dialect('unixpwd', delimiter=':', quoting=csv.QUOTE_NONE)
Eli Benderskyec40bab2011-03-13 08:45:19 +0200509 with open('passwd', 'rb') as f:
510 reader = csv.reader(f, 'unixpwd')
Georg Brandl8ec7f652007-08-15 14:28:01 +0000511
512A slightly more advanced use of the reader --- catching and reporting errors::
513
Benjamin Petersona7b55a32009-02-20 03:31:23 +0000514 import csv, sys
Eli Benderskyec40bab2011-03-13 08:45:19 +0200515 filename = 'some.csv'
516 with open(filename, 'rb') as f:
517 reader = csv.reader(f)
518 try:
519 for row in reader:
520 print row
Andrew Svetlov1625d882012-10-30 21:56:43 +0200521 except csv.Error as e:
Eli Benderskyec40bab2011-03-13 08:45:19 +0200522 sys.exit('file %s, line %d: %s' % (filename, reader.line_num, e))
Georg Brandl8ec7f652007-08-15 14:28:01 +0000523
524And while the module doesn't directly support parsing strings, it can easily be
525done::
526
527 import csv
528 for row in csv.reader(['one,two,three']):
529 print row
530
531The :mod:`csv` module doesn't directly support reading and writing Unicode, but
532it is 8-bit-clean save for some problems with ASCII NUL characters. So you can
533write functions or classes that handle the encoding and decoding for you as long
534as you avoid encodings like UTF-16 that use NULs. UTF-8 is recommended.
535
Georg Brandlcf3fb252007-10-21 10:52:38 +0000536:func:`unicode_csv_reader` below is a :term:`generator` that wraps :class:`csv.reader`
Georg Brandl8ec7f652007-08-15 14:28:01 +0000537to handle Unicode CSV data (a list of Unicode strings). :func:`utf_8_encoder`
Georg Brandlcf3fb252007-10-21 10:52:38 +0000538is a :term:`generator` that encodes the Unicode strings as UTF-8, one string (or row) at
Georg Brandl8ec7f652007-08-15 14:28:01 +0000539a time. The encoded strings are parsed by the CSV reader, and
540:func:`unicode_csv_reader` decodes the UTF-8-encoded cells back into Unicode::
541
542 import csv
543
544 def unicode_csv_reader(unicode_csv_data, dialect=csv.excel, **kwargs):
545 # csv.py doesn't do Unicode; encode temporarily as UTF-8:
546 csv_reader = csv.reader(utf_8_encoder(unicode_csv_data),
547 dialect=dialect, **kwargs)
548 for row in csv_reader:
549 # decode UTF-8 back to Unicode, cell by cell:
550 yield [unicode(cell, 'utf-8') for cell in row]
551
552 def utf_8_encoder(unicode_csv_data):
553 for line in unicode_csv_data:
554 yield line.encode('utf-8')
555
556For all other encodings the following :class:`UnicodeReader` and
557:class:`UnicodeWriter` classes can be used. They take an additional *encoding*
558parameter in their constructor and make sure that the data passes the real
559reader or writer encoded as UTF-8::
560
Benjamin Petersona7b55a32009-02-20 03:31:23 +0000561 import csv, codecs, cStringIO
Georg Brandl8ec7f652007-08-15 14:28:01 +0000562
563 class UTF8Recoder:
564 """
565 Iterator that reads an encoded stream and reencodes the input to UTF-8
566 """
567 def __init__(self, f, encoding):
568 self.reader = codecs.getreader(encoding)(f)
569
570 def __iter__(self):
571 return self
572
573 def next(self):
574 return self.reader.next().encode("utf-8")
575
576 class UnicodeReader:
577 """
578 A CSV reader which will iterate over lines in the CSV file "f",
579 which is encoded in the given encoding.
580 """
581
582 def __init__(self, f, dialect=csv.excel, encoding="utf-8", **kwds):
583 f = UTF8Recoder(f, encoding)
584 self.reader = csv.reader(f, dialect=dialect, **kwds)
585
586 def next(self):
587 row = self.reader.next()
588 return [unicode(s, "utf-8") for s in row]
589
590 def __iter__(self):
591 return self
592
593 class UnicodeWriter:
594 """
595 A CSV writer which will write rows to CSV file "f",
596 which is encoded in the given encoding.
597 """
598
599 def __init__(self, f, dialect=csv.excel, encoding="utf-8", **kwds):
600 # Redirect output to a queue
601 self.queue = cStringIO.StringIO()
602 self.writer = csv.writer(self.queue, dialect=dialect, **kwds)
603 self.stream = f
604 self.encoder = codecs.getincrementalencoder(encoding)()
605
606 def writerow(self, row):
607 self.writer.writerow([s.encode("utf-8") for s in row])
608 # Fetch UTF-8 output from the queue ...
609 data = self.queue.getvalue()
610 data = data.decode("utf-8")
611 # ... and reencode it into the target encoding
612 data = self.encoder.encode(data)
613 # write to the target stream
614 self.stream.write(data)
615 # empty queue
616 self.queue.truncate(0)
617
618 def writerows(self, rows):
619 for row in rows:
620 self.writerow(row)
621