pandas-should

pandas-should is pandas extension for asserting.

Install

Install and update released version using pip:

$ pip install pandas-should

If you want development version:

$ pip install git+https://github.com/momijiame/pandas-should.git

Quickstart

You can just to import pandas_should to use:

import pandas_should

DataFrame

Length (rows)

Before:

assert len(df1) == len(df2)

After:

assert df1.should.have_same_length(df2)

Or

assert df1.should.have_length(len(df2))

Before (multiple DataFrame):

assert len(df1) == (len(df2) + len(df3))

After (multiple DataFrame)

assert df1.should.have_same_length(df2, df3)

Width (columns)

Before:

assert len(df1.columns) == len(df2.columns)

After:

assert df1.should.have_same_width(df2)

Or

assert df1.should.have_width(len(df2.columns))

Before (multiple DataFrame):

assert len(df1.columns) == (len(df2.columns)) + len(df3.columns)))

After (multiple DataFrame)

assert df1.should.have_same_width(df2, df3)

Equality

Before:

from pandas.util.testing import assert_frame_equal
def equal(a, b):
    try:
        assert_frame_equal(a, b)
    except AssertionError:
        return False
    return True

assert equal(df1, df2)

After:

assert df1.should.equal(df2)

Null inclusion

Before:

assert df.isnull().any(axis=None)

After:

assert df.should.have_null()

Or expect not to be included:

assert df.should.have_not_null()
assert not df.should.have_null()

Shape

Before:

assert df1.shape == df2.shape

After:

assert df1.should.be_shaped_like(df2)

Or

assert df1.should.be_shaped_like(df2.shape)
assert df1.should.be_shaped_like(df2.shape[0], df2.shape[1])

Value range

Before:

assert (df >= range_min).any(axis=None) and (df <= range_max).any(axis=None) 

After:

assert df.should.fall_within_range(range_min, range_max)

Greater than only:

assert df.should.greater_than(range_min)
assert df.should.greater_than_or_equal(range_min)

Less than only:

assert df.should.less_than(range_min)
assert df.should.less_than_or_equal(range_min)

Series

Length

Before:

assert len(s1) == len(s2)

After:

assert s1.should.have_same_length(s2)

Or:

assert s1.should.have_length(len(s2))

Before (multiple Series):

assert len(s1) == (len(s2) + len(s3))

After (multiple Series):

assert s1.should.have_same_length(s2, s3)

Equality

Before:

from pandas.util.testing import assert_series_equal
def equal(a, b):
    try:
        assert_series_equal(a, b)
    except AssertionError:
        return False
    return True

assert equal(s1, s2)

After:

assert s1.should.equal(s2)

Null inclusion

Before:

assert s.isnull().any()

After:

assert s.should.have_null()

Or expect not to be included:

assert s.should.have_not_null()
assert not s.should.have_null()

Value range

Before:

assert (s >= range_min).any() and (s <= range_max).any() 

After:

assert s.should.fall_within_range(range_min, range_max)

Greater than only:

assert s.should.greater_than(range_min)
assert s.should.greater_than_or_equal(range_min)

Less than only:

assert s.should.less_than(range_min)
assert s.should.less_than_or_equal(range_min)

Value variety

Before:

assert len(s.unique()) == expect_size

After:

assert s.should.have_number_of_unique_values(expect_size)