Difference between revisions of "The configparser module in Python"
m |
m |
||
| Line 1: | Line 1: | ||
| − | The official documentation for the configparser module can be found [https://docs.python.org/3/library/configparser.html here]. The current page is only meant as my comments to that page. For actual proper documentation of the module, please refer to official page. | + | The official documentation for the configparser module can be found [https://docs.python.org/3/library/configparser.html here]. The current page is only meant as my comments to that page. For actual proper documentation of the module, please refer to official page. Unles otherwise specified all discussion on this page assumes default values of all options for the ConfigParser object. |
==DEFAULT and other section names== | ==DEFAULT and other section names== | ||
| Line 24: | Line 24: | ||
Assuming the <tt>DEFAULT</tt> section, when given en empty string, is somewhat documented in the discussion of the <tt>has_option(section, option)</tt> method, where it states "If the specified section is None or an empty string, DEFAULT is assumed." But leaving an empty section called ''[empty-string]'' isn't mentioned at all. Leaving this empty section in the written file breaks configparser itself, because it will throw an exception, when reading the file containing this empty section. | Assuming the <tt>DEFAULT</tt> section, when given en empty string, is somewhat documented in the discussion of the <tt>has_option(section, option)</tt> method, where it states "If the specified section is None or an empty string, DEFAULT is assumed." But leaving an empty section called ''[empty-string]'' isn't mentioned at all. Leaving this empty section in the written file breaks configparser itself, because it will throw an exception, when reading the file containing this empty section. | ||
| − | Keeping the test.ini file written by the earlier code, we can use this code to read it: | + | Keeping the <tt>test.ini</tt> file written by the earlier code, we can use this code to read it: |
import configparser as cp | import configparser as cp | ||
conf = cp.ConfigParser() | conf = cp.ConfigParser() | ||
| Line 32: | Line 32: | ||
configparser.ParsingError: Source contains parsing errors: 'test.ini' | configparser.ParsingError: Source contains parsing errors: 'test.ini' | ||
[line 6]: '[]\n' | [line 6]: '[]\n' | ||
| + | This seems to be because the section is empty, since putting a value in the section, i.e. like this | ||
| + | [] | ||
| + | float value 1 = 3.1415 | ||
| + | |||
| + | will produce this exception: | ||
| + | configparser.DuplicateOptionError: While reading from 'test.ini' [line 7]: option 'float value 1' in section 'DEFAULT' already exists | ||
| + | |||
| + | Specifying <tt>strict=False</tt> when creating the object will revert this exception to the previous (<tt> configparser.ParsingError: Source contains parsing errors: 'test.ini'\n[line 6]: '[]\n'</tt>). | ||
[[Category:Programmering]][[Category:Sprog]][[Category:Python]] | [[Category:Programmering]][[Category:Sprog]][[Category:Python]] | ||
Revision as of 14:07, 20 April 2022
The official documentation for the configparser module can be found here. The current page is only meant as my comments to that page. For actual proper documentation of the module, please refer to official page. Unles otherwise specified all discussion on this page assumes default values of all options for the ConfigParser object.
DEFAULT and other section names
The official documentation mentions that default values will override fallback values, if the default value exist when get is called. But if you aren't reading the documentation serially, then you might not know, that default values is taken from the section named 'DEFAULT'.
As of the time of writing this (2022-04-20), if you create a section named by the empty string, it will display as precisely that (a section called [empty-string] containing whatever key.value pairs is put in it), until the file is written to disk. When the file is written to disk, the key,value pairs in the section called [empty-string] will be written to the DEFAULT section, and an empty section, called [empty-string], will be created.
This means that this code:
import configparser as cp
conf = cp.ConfigParser()
conf[] = {'Int value 1': 42, 'String value 1': 'String value test', 'Float value 1': 3.1415}
with open('test.ini', 'w') as f:
conf.write(f)
Will produce this file:
[DEFAULT] int value 1 = 42 string value 1 = String value test float value 1 = 3.1415 []
Assuming the DEFAULT section, when given en empty string, is somewhat documented in the discussion of the has_option(section, option) method, where it states "If the specified section is None or an empty string, DEFAULT is assumed." But leaving an empty section called [empty-string] isn't mentioned at all. Leaving this empty section in the written file breaks configparser itself, because it will throw an exception, when reading the file containing this empty section.
Keeping the test.ini file written by the earlier code, we can use this code to read it:
import configparser as cp
conf = cp.ConfigParser()
conf.read('test.ini')
This will produce an exception, caused by the empty section:
configparser.ParsingError: Source contains parsing errors: 'test.ini'
[line 6]: '[]\n'
This seems to be because the section is empty, since putting a value in the section, i.e. like this
[] float value 1 = 3.1415
will produce this exception:
configparser.DuplicateOptionError: While reading from 'test.ini' [line 7]: option 'float value 1' in section 'DEFAULT' already exists
Specifying strict=False when creating the object will revert this exception to the previous ( configparser.ParsingError: Source contains parsing errors: 'test.ini'\n[line 6]: '[]\n').