def test_recast_1():
table = (('id', 'variable', 'value'),
(3, 'age', 16),
(1, 'gender', 'F'),
(2, 'gender', 'M'),
(2, 'age', 17),
(1, 'age', 12),
(3, 'gender', 'M'))
expectation = (('id', 'age', 'gender'),
(1, 12, 'F'),
(2, 17, 'M'),
(3, 16, 'M'))
# by default lift 'variable' field, hold everything else
result = recast(table)
ieq(expectation, result)
result = recast(table, variablefield='variable')
ieq(expectation, result)
result = recast(table, key='id', variablefield='variable')
ieq(expectation, result)
result = recast(table, key='id', variablefield='variable',
valuefield='value')
ieq(expectation, result)
def test_recast_date():
dt = datetime.now().replace
table = (('id', 'variable', 'value'),
(dt(hour=3), 'age', 16),
(dt(hour=1), 'gender', 'F'),
(dt(hour=2), 'gender', 'M'),
(dt(hour=2), 'age', 17),
(dt(hour=1), 'age', 12),
(dt(hour=3), 'gender', 'M'))
expectation = (('id', 'age', 'gender'),
(dt(hour=1), 12, 'F'),
(dt(hour=2), 17, 'M'),
(dt(hour=3), 16, 'M'))
# by default lift 'variable' field, hold everything else
result = recast(table)
ieq(expectation, result)
result = recast(table, variablefield='variable')
ieq(expectation, result)
result = recast(table, key='id', variablefield='variable')
ieq(expectation, result)
result = recast(table, key='id', variablefield='variable',
valuefield='value')
ieq(expectation, result)
def test_recast4():
# deal with missing data
table = (('id', 'variable', 'value'), (1, 'gender', 'F'), (2, 'age', 17),
(1, 'age', 12), (3, 'gender', 'M'))
result = recast(table, key='id')
expect = (('id', 'age', 'gender'), (1, 12, 'F'), (2, 17, None), (3, None,
'M'))
ieq(expect, result)
def test_recast_3():
table = (('id', 'time', 'variable', 'value'), (1, 11, 'weight', 66.4),
(1, 14, 'weight', 55.2), (2, 12, 'weight', 53.2),
(2, 16, 'weight', 43.3), (3, 12, 'weight',
34.5), (3, 17, 'weight', 49.4))
expectation = (('id', 'time', 'weight'), (1, 11, 66.4), (1, 14, 55.2),
(2, 12, 53.2), (2, 16, 43.3), (3, 12, 34.5), (3, 17, 49.4))
result = recast(table)
ieq(expectation, result)
# in the absence of an aggregation function, list all values
expectation = (('id', 'weight'), (1, [66.4, 55.2]), (2, [53.2, 43.3]),
(3, [34.5, 49.4]))
result = recast(table, key='id')
ieq(expectation, result)
# max aggregation
expectation = (('id', 'weight'), (1, 66.4), (2, 53.2), (3, 49.4))
result = recast(table, key='id', reducers={'weight': max})
ieq(expectation, result)
# min aggregation
expectation = (('id', 'weight'), (1, 55.2), (2, 43.3), (3, 34.5))
result = recast(table, key='id', reducers={'weight': min})
ieq(expectation, result)
# mean aggregation
expectation = (('id', 'weight'), (1, 60.80), (2, 48.25), (3, 41.95))
def mean(values):
return float(sum(values)) / len(values)
def meanf(precision):
def f(values):
v = mean(values)
v = round(v, precision)
return v
return f
result = recast(table, key='id', reducers={'weight': meanf(precision=2)})
ieq(expectation, result)
def test_recast_2():
table = (('id', 'variable', 'value'), (3, 'age', 16), (1, 'gender', 'F'),
(2, 'gender', 'M'), (2, 'age', 17), (1, 'age', 12), (3, 'gender',
'M'))
expectation = (('id', 'gender'), (1, 'F'), (2, 'M'), (3, 'M'))
# can manually pick which variables you want to recast as fields
# TODO this is awkward
result = recast(table, key='id', variablefield={'variable': ['gender']})
ieq(expectation, result)
def test_recast4():
# deal with missing data
table = (('id', 'variable', 'value'),
(1, 'gender', 'F'),
(2, 'age', 17),
(1, 'age', 12),
(3, 'gender', 'M'))
result = recast(table, key='id')
expect = (('id', 'age', 'gender'),
(1, 12, 'F'),
(2, 17, None),
(3, None, 'M'))
ieq(expect, result)
def test_recast_2():
table = (('id', 'variable', 'value'),
(3, 'age', 16),
(1, 'gender', 'F'),
(2, 'gender', 'M'),
(2, 'age', 17),
(1, 'age', 12),
(3, 'gender', 'M'))
expectation = (('id', 'gender'),
(1, 'F'),
(2, 'M'),
(3, 'M'))
# can manually pick which variables you want to recast as fields
# TODO this is awkward
result = recast(table, key='id', variablefield={'variable': ['gender']})
ieq(expectation, result)
def test_recast_empty():
table = (('foo', 'variable', 'value'),)
expect = (('foo',),)
actual = recast(table)
ieq(expect, actual)
def test_recast_3():
table = (('id', 'time', 'variable', 'value'),
(1, 11, 'weight', 66.4),
(1, 14, 'weight', 55.2),
(2, 12, 'weight', 53.2),
(2, 16, 'weight', 43.3),
(3, 12, 'weight', 34.5),
(3, 17, 'weight', 49.4))
expectation = (('id', 'time', 'weight'),
(1, 11, 66.4),
(1, 14, 55.2),
(2, 12, 53.2),
(2, 16, 43.3),
(3, 12, 34.5),
(3, 17, 49.4))
result = recast(table)
ieq(expectation, result)
# in the absence of an aggregation function, list all values
expectation = (('id', 'weight'),
(1, [66.4, 55.2]),
(2, [53.2, 43.3]),
(3, [34.5, 49.4]))
result = recast(table, key='id')
ieq(expectation, result)
# max aggregation
expectation = (('id', 'weight'),
(1, 66.4),
(2, 53.2),
(3, 49.4))
result = recast(table, key='id', reducers={'weight': max})
ieq(expectation, result)
# min aggregation
expectation = (('id', 'weight'),
(1, 55.2),
(2, 43.3),
(3, 34.5))
result = recast(table, key='id', reducers={'weight': min})
ieq(expectation, result)
# mean aggregation
expectation = (('id', 'weight'),
(1, 60.80),
(2, 48.25),
(3, 41.95))
def mean(values):
return float(sum(values)) / len(values)
def meanf(precision):
def f(values):
v = mean(values)
v = round(v, precision)
return v
return f
result = recast(table, key='id', reducers={'weight': meanf(precision=2)})
ieq(expectation, result)