472ebf9c88f2c21b94e0ec2b377957f1274d5fe7,tests/test_standardise.py,TestStandardise,test_clean,#TestStandardise#,44

Before Change


        expected = ["mary ann", "bob", "angel", "bob", "mary ann", "john", np.nan]

        s = StandardSeries(values)
        s = s.clean(lower=True, replace_by_none="[^ \-\_A-Za-z0-9]+", replace_by_whitespace="[\-\_]", remove_brackets=True, inplace=False)

        s_exp = StandardSeries(expected)

        pdt.assert_series_equal(s, s_exp)

After Change



    def test_clean(self):

        values = pd.Series(["Mary-ann", "Bob :)", "Angel", "Bob (alias Billy)", "Mary  ann", "John", np.nan])
        expected = pd.Series(["mary ann", "bob", "angel", "bob", "mary ann", "john", np.nan])

        clean_series = recordlinkage.clean(values)

        // Check if series are identical.
Italian Trulli
In pattern: SUPERPATTERN

Frequency: 3

Non-data size: 3

Instances


Project Name: J535D165/recordlinkage
Commit Name: 472ebf9c88f2c21b94e0ec2b377957f1274d5fe7
Time: 2016-02-06
Author: jonathandebruinhome@gmail.com
File Name: tests/test_standardise.py
Class Name: TestStandardise
Method Name: test_clean


Project Name: J535D165/recordlinkage
Commit Name: 472ebf9c88f2c21b94e0ec2b377957f1274d5fe7
Time: 2016-02-06
Author: jonathandebruinhome@gmail.com
File Name: tests/test_standardise.py
Class Name: TestStandardise
Method Name: test_clean


Project Name: J535D165/recordlinkage
Commit Name: 472ebf9c88f2c21b94e0ec2b377957f1274d5fe7
Time: 2016-02-06
Author: jonathandebruinhome@gmail.com
File Name: tests/test_standardise.py
Class Name: TestStandardise
Method Name: test_clean_lower


Project Name: J535D165/recordlinkage
Commit Name: 472ebf9c88f2c21b94e0ec2b377957f1274d5fe7
Time: 2016-02-06
Author: jonathandebruinhome@gmail.com
File Name: tests/test_standardise.py
Class Name: TestStandardise
Method Name: test_clean_brackets