junitparser handles JUnit/xUnit Result XML files. Use it to parse and manipulate existing Result XML files, or create new JUnit/xUnit result XMLs from scratch.
- Parse or modify existing JUnit/xUnit XML files.
- Parse or modify non-standard or customized JUnit/xUnit XML files, by monkey patching existing element definitions.
- Create JUnit/xUnit test results from scratch.
- Merge test result XML files.
- Specify XML parser. For example you can use lxml to speed things up.
- Invoke from command line, or python -m junitparser
- Python 2 and 3 support (As of Nov 2020, 1/4 of the users are still on Python 2, so there is no plan to drop Python 2 support)
Version 2 improved support for pytest result XML files by fixing a few issues,
notably that there could be multiple <Failure> or <Error> entries. There is a
breaking change that TestCase.result
is now a list instead of a single item.
If you are using this attribute, please update your code accordingly.
pip install junitparser
You should be relatively familiar with the Junit XML format. If not, run
pydoc
on the exposed classes and functions to see how it's structured.
You have some test result data, and you want to convert them into junit.xml format.
from junitparser import TestCase, TestSuite, JUnitXml, Skipped, Error
# Create cases
case1 = TestCase('case1', 'class.name', 0.5) # params are optional
case1.classname = "modified.class.name" # specify or change case attrs
case1.result = [Skipped()] # You can have a list of results
case2 = TestCase('case2')
case2.result = [Error('Example error message', 'the_error_type')]
# Create suite and add cases
suite = TestSuite('suite1')
suite.add_property('build', '55')
suite.add_testcase(case1)
suite.add_testcase(case2)
suite.remove_testcase(case2)
#Bulk add cases to suite
case3 = TestCase('case3')
case4 = TestCase('case4')
suite.add_testcases([case3, case4])
# Add suite to JunitXml
xml = JUnitXml()
xml.add_testsuite(suite)
xml.write('junit.xml')
You have some existing junit.xml files, and you want to modify the content.
from junitparser import JUnitXml
xml = JUnitXml.fromfile('/path/to/junit.xml')
for suite in xml:
# handle suites
for case in suite:
# handle cases
xml.write() # Writes back to file
It is also possible to use a custom parser. For example lxml provides a plethora of parsing options. We can use them this way:
from lxml.etree import XMLParser, parse
from junitparser import JUnitXml
def parse_func(file_path):
xml_parser = XMLParser(huge_tree=True)
return parse(file_path, xml_parser)
xml = JUnitXml.fromfile('/path/to/junit.xml', parse_func)
# process xml...
You have two or more XML files, and you want to merge them into one.
from junitparser import JUnitXml
xml1 = JUnitXml.fromfile('/path/to/junit1.xml')
xml2 = JUnitXml.fromfile('/path/to/junit2.xml')
newxml = xml1 + xml2
# Alternatively, merge in place
xml1 += xml2
Note that it won't check for duplicate entries. You need to deal with them on your own.
By default junitparser supports the schema of windyroad, which is a relatively simple schema.
Junitparser also support extra schemas:
from junitparser.xunit2 import JUnitParser, TestCase, TestSuite, \
RerunFailure
# These classes are redefined to support extra properties and attributes
# of the xunit2 schema.
suite = TestSuite("mySuite")
suite.system_err = "System err" # xunit2 specific property
case = TestCase("myCase")
rerun_failure = RerunFailure("Not found", "404") # case property
rerun_failure.stack_trace = "Stack"
rerun_failure.system_err = "E404"
rerun_failure.system_out = "NOT FOUND"
case.add_rerun_result(rerun_failure)
Currently supported schemas including:
- xunit2, supported by pytest, Erlang/OTP, Maven Surefire, CppTest, etc.
PRs are welcome to support more schemas.
You want to use an attribute that is not supported by default.
from junitparser import TestCase, Attr, IntAttr, FloatAttr
# Add the custom attribute
TestCase.id = IntAttr('id')
TestCase.rate = FloatAttr('rate')
TestCase.custom = Attr('custom')
case = TestCase()
case.id = 123
case.rate = 0.95
case.custom = 'foobar'
There may be once in 1000 years you want to it this way, but anyways. Suppose you want to add element CustomElement to TestCase.
from junitparser import Element, Attr, TestSuite
# Create the new element by subclassing Element,
# and add custom attributes to it.
class CustomElement(Element):
_tag = 'custom'
foo = Attr()
bar = Attr()
testcase = TestCase()
custom = CustomElement()
testcase.append(custom)
# To find a single sub-element:
testcase.child(CustomElement)
# To iterate over custom elements:
for custom in testcase.iterchildren(CustomElement):
... # Do things with custom element
Say you have some data stored in the XML as custom attributes and you want to read them out:
from junitparser import TestCase, Attr, JUnitXml
# Create the new element by subclassing Element or one of its child class,
# and add custom attributes to it.
class MyTestCase(TestCase):
foo = Attr()
xml = JUnitXml.fromfile('/path/to/junit.xml')
for suite in xml:
# handle suites
for case in suite:
my_case = MyTestCase.fromelem(case)
print(my_case.foo)
$ junitparser --help
usage: junitparser [-h] [-v] {merge} ...
Junitparser CLI helper.
positional arguments:
{merge} command
merge Merge Junit XML format reports with junitparser.
verify Return a non-zero exit code if one of the testcases failed or errored.
optional arguments:
-h, --help show this help message and exit
-v, --version show program's version number and exit
$ junitparser merge --help
usage: junitparser merge [-h] [--glob] paths [paths ...] output
positional arguments:
paths Original XML path(s).
output Merged XML Path, setting to "-" will output console
optional arguments:
-h, --help show this help message and exit
--glob Treat original XML path(s) as glob(s).
--suite-name SUITE_NAME
Name added to <testsuites>.
$ junitparser verify --help
usage: junitparser verify [-h] [--glob] paths [paths ...]
positional arguments:
paths XML path(s) of reports to verify.
optional arguments:
-h, --help show this help message and exit
--glob Treat original XML path(s) as glob(s).
The tests are written with python unittest
, to run them, use pytest:
pytest test.py
PRs are welcome!