def test_fwf_colspecs_None(self):
# GH 7079
data = """\
123456
456789
"""
colspecs = [(0, 3), (3, None)]
result = read_fwf(StringIO(data), colspecs=colspecs, header=None)
expected = DataFrame([[123, 456], [456, 789]])
tm.assert_frame_equal(result, expected)
colspecs = [(None, 3), (3, 6)]
result = read_fwf(StringIO(data), colspecs=colspecs, header=None)
expected = DataFrame([[123, 456], [456, 789]])
tm.assert_frame_equal(result, expected)
colspecs = [(0, None), (3, None)]
result = read_fwf(StringIO(data), colspecs=colspecs, header=None)
expected = DataFrame([[123456, 456], [456789, 789]])
tm.assert_frame_equal(result, expected)
colspecs = [(None, None), (3, 6)]
result = read_fwf(StringIO(data), colspecs=colspecs, header=None)
expected = DataFrame([[123456, 456], [456789, 789]])
tm.assert_frame_equal(result, expected)
def test_skiprows_inference_empty(self):
test = """
AA BBB C
12 345 6
78 901 2
""".strip()
with pytest.raises(EmptyDataError):
read_fwf(StringIO(test), skiprows=3)
def test_skiprows_inference_empty():
data = """
AA BBB C
12 345 6
78 901 2
""".strip()
msg = "No rows from which to infer column width"
with pytest.raises(EmptyDataError, match=msg):
read_fwf(StringIO(data), skiprows=3)
def test_under_specified():
data = """\
A B C D E
201158 360.242940 149.910199 11950.7
201159 444.953632 166.985655 11788.4
201160 364.136849 183.628767 11806.2
201161 413.836124 184.375703 11916.8
201162 502.953953 173.237159 12468.3
"""
with pytest.raises(ValueError, match="Must specify either"):
read_fwf(StringIO(data), colspecs=None, widths=None)
def test_bool_header_arg(self):
# see gh-6114
data = """\
MyColumn
a
b
a
b"""
for arg in [True, False]:
with pytest.raises(TypeError):
read_fwf(StringIO(data), header=arg)
def test_bool_header_arg(header):
# see gh-6114
data = """\
MyColumn
a
b
a
b"""
msg = "Passing a bool to header is invalid"
with pytest.raises(TypeError, match=msg):
read_fwf(StringIO(data), header=header)
def test_variable_width_unicode():
data = """
שלום שלום
ום שלל
של ום
""".strip("\r\n")
encoding = "utf8"
kwargs = dict(header=None, encoding=encoding)
expected = read_fwf(BytesIO(data.encode(encoding)),
colspecs=[(0, 4), (5, 9)], **kwargs)
result = read_fwf(BytesIO(data.encode(encoding)), **kwargs)
tm.assert_frame_equal(result, expected)
def test_full_file(self):
# File with all values
test = '''index A B C
2000-01-03T00:00:00 0.980268513777 3 foo
2000-01-04T00:00:00 1.04791624281 -4 bar
2000-01-05T00:00:00 0.498580885705 73 baz
2000-01-06T00:00:00 1.12020151869 1 foo
2000-01-07T00:00:00 0.487094399463 0 bar
2000-01-10T00:00:00 0.836648671666 2 baz
2000-01-11T00:00:00 0.157160753327 34 foo'''
colspecs = ((0, 19), (21, 35), (38, 40), (42, 45))
expected = read_fwf(StringIO(test), colspecs=colspecs)
tm.assert_frame_equal(expected, read_fwf(StringIO(test)))
def test_full_file_with_spaces(self):
# File with spaces in columns
test = """
Account Name Balance CreditLimit AccountCreated
101 Keanu Reeves 9315.45 10000.00 1/17/1998
312 Gerard Butler 90.00 1000.00 8/6/2003
868 Jennifer Love Hewitt 0 17000.00 5/25/1985
761 Jada Pinkett-Smith 49654.87 100000.00 12/5/2006
317 Bill Murray 789.65 5000.00 2/5/2007
""".strip('\r\n')
colspecs = ((0, 7), (8, 28), (30, 38), (42, 53), (56, 70))
expected = read_fwf(StringIO(test), colspecs=colspecs)
tm.assert_frame_equal(expected, read_fwf(StringIO(test)))
def test_full_file_with_spaces_and_missing(self):
# File with spaces and missing values in columsn
test = """
Account Name Balance CreditLimit AccountCreated
101 10000.00 1/17/1998
312 Gerard Butler 90.00 1000.00 8/6/2003
868 5/25/1985
761 Jada Pinkett-Smith 49654.87 100000.00 12/5/2006
317 Bill Murray 789.65
""".strip('\r\n')
colspecs = ((0, 7), (8, 28), (30, 38), (42, 53), (56, 70))
expected = read_fwf(StringIO(test), colspecs=colspecs)
tm.assert_frame_equal(expected, read_fwf(StringIO(test)))
def test_messed_up_data(self):
# Completely messed up file
test = """
Account Name Balance Credit Limit Account Created
101 10000.00 1/17/1998
312 Gerard Butler 90.00 1000.00
761 Jada Pinkett-Smith 49654.87 100000.00 12/5/2006
317 Bill Murray 789.65
""".strip('\r\n')
colspecs = ((2, 10), (15, 33), (37, 45), (49, 61), (64, 79))
expected = read_fwf(StringIO(test), colspecs=colspecs)
tm.assert_frame_equal(expected, read_fwf(StringIO(test)))
def test_over_specified():
data = """\
A B C D E
201158 360.242940 149.910199 11950.7
201159 444.953632 166.985655 11788.4
201160 364.136849 183.628767 11806.2
201161 413.836124 184.375703 11916.8
201162 502.953953 173.237159 12468.3
"""
colspecs = [(0, 4), (4, 8), (8, 20), (21, 33), (34, 43)]
with pytest.raises(ValueError, match="must specify only one of"):
read_fwf(StringIO(data), colspecs=colspecs, widths=[6, 10, 10, 7])
def test_fwf_colspecs_is_list_or_tuple_of_two_element_tuples(self):
data = """index,A,B,C,D
foo,2,3,4,5
bar,7,8,9,10
baz,12,13,14,15
qux,12,13,14,15
foo2,12,13,14,15
bar2,12,13,14,15
"""
with tm.assertRaisesRegexp(TypeError,
'Each column specification must be.+'):
read_fwf(StringIO(data), [('a', 1)])
def test_full_file_with_missing(self):
# File with missing values
test = """index A B C
2000-01-03T00:00:00 0.980268513777 3 foo
2000-01-04T00:00:00 1.04791624281 -4 bar
0.498580885705 73 baz
2000-01-06T00:00:00 1.12020151869 1 foo
2000-01-07T00:00:00 0 bar
2000-01-10T00:00:00 0.836648671666 2 baz
34"""
colspecs = ((0, 19), (21, 35), (38, 40), (42, 45))
expected = read_fwf(StringIO(test), colspecs=colspecs)
tm.assert_frame_equal(expected, read_fwf(StringIO(test)))
def test_fwf_colspecs_is_list_or_tuple_of_two_element_tuples(self):
data = """index,A,B,C,D
foo,2,3,4,5
bar,7,8,9,10
baz,12,13,14,15
qux,12,13,14,15
foo2,12,13,14,15
bar2,12,13,14,15
"""
msg = 'Each column specification must be.+'
with pytest.raises(TypeError, match=msg):
read_fwf(StringIO(data), [('a', 1)])
def test_variable_width_unicode(self):
if not compat.PY3:
pytest.skip(
'Bytes-related test - only needs to work on Python 3')
test = """
שלום שלום
ום שלל
של ום
""".strip('\r\n')
expected = read_fwf(BytesIO(test.encode('utf8')),
colspecs=[(0, 4), (5, 9)],
header=None, encoding='utf8')
tm.assert_frame_equal(expected, read_fwf(
BytesIO(test.encode('utf8')), header=None, encoding='utf8'))
def test_multiple_delimiters(self):
test = r"""
col1~~~~~col2 col3++++++++++++++++++col4
~~22.....11.0+++foo~~~~~~~~~~Keanu Reeves
33+++122.33\\\bar.........Gerard Butler
++44~~~~12.01 baz~~Jennifer Love Hewitt
~~55 11+++foo++++Jada Pinkett-Smith
..66++++++.03~~~bar Bill Murray
""".strip('\r\n')
colspecs = ((0, 4), (7, 13), (15, 19), (21, 41))
expected = read_fwf(StringIO(test), colspecs=colspecs,
delimiter=' +~.\\')
tm.assert_frame_equal(expected, read_fwf(StringIO(test),
delimiter=' +~.\\'))
def test_fwf_colspecs_is_list_or_tuple():
data = """index,A,B,C,D
foo,2,3,4,5
bar,7,8,9,10
baz,12,13,14,15
qux,12,13,14,15
foo2,12,13,14,15
bar2,12,13,14,15
"""
msg = "column specifications must be a list or tuple.+"
with pytest.raises(TypeError, match=msg):
read_fwf(StringIO(data), colspecs={"a": 1}, delimiter=",")
def test_fwf_colspecs_infer_nrows(self):
# GH 15138
data = """\
1 2
123 98
"""
# infer_nrows == 1 should have colspec == [(2, 3), (5, 6)]
df = read_fwf(StringIO(data), header=None, infer_nrows=1)
expected = pd.DataFrame([[1, 2], [3, 8]])
tm.assert_frame_equal(df, expected)
# test for infer_nrows > number of rows
df = read_fwf(StringIO(data), header=None, infer_nrows=10)
expected = pd.DataFrame([[1, 2], [123, 98]])
tm.assert_frame_equal(df, expected)
def test_fwf_regression(self):
# GH 3594
# turns out 'T060' is parsable as a datetime slice!
tzlist = [1, 10, 20, 30, 60, 80, 100]
ntz = len(tzlist)
tcolspecs = [16] + [8] * ntz
tcolnames = ['SST'] + ["T%03d" % z for z in tzlist[1:]]
data = """ 2009164202000 9.5403 9.4105 8.6571 7.8372 6.0612 5.8843 5.5192
2009164203000 9.5435 9.2010 8.6167 7.8176 6.0804 5.8728 5.4869
2009164204000 9.5873 9.1326 8.4694 7.5889 6.0422 5.8526 5.4657
2009164205000 9.5810 9.0896 8.4009 7.4652 6.0322 5.8189 5.4379
2009164210000 9.6034 9.0897 8.3822 7.4905 6.0908 5.7904 5.4039
"""
df = read_fwf(StringIO(data),
index_col=0,
header=None,
names=tcolnames,
widths=tcolspecs,
parse_dates=True,
date_parser=lambda s: datetime.strptime(s, '%Y%j%H%M%S'))
for c in df.columns:
res = df.loc[:, c]
assert len(res)
def test_non_space_filler():
# From Thomas Kluyver:
#
# Apparently, some non-space filler characters can be seen, this is
# supported by specifying the 'delimiter' character:
#
# http://publib.boulder.ibm.com/infocenter/dmndhelp/v6r1mx/index.jsp?topic=/com.ibm.wbit.612.help.config.doc/topics/rfixwidth.html
data = """\
A~~~~B~~~~C~~~~~~~~~~~~D~~~~~~~~~~~~E
201158~~~~360.242940~~~149.910199~~~11950.7
201159~~~~444.953632~~~166.985655~~~11788.4
201160~~~~364.136849~~~183.628767~~~11806.2
201161~~~~413.836124~~~184.375703~~~11916.8
201162~~~~502.953953~~~173.237159~~~12468.3
"""
colspecs = [(0, 4), (4, 8), (8, 20), (21, 33), (34, 43)]
result = read_fwf(StringIO(data), colspecs=colspecs, delimiter="~")
expected = DataFrame([[2011, 58, 360.242940, 149.910199, 11950.7],
[2011, 59, 444.953632, 166.985655, 11788.4],
[2011, 60, 364.136849, 183.628767, 11806.2],
[2011, 61, 413.836124, 184.375703, 11916.8],
[2011, 62, 502.953953, 173.237159, 12468.3]],
columns=["A", "B", "C", "D", "E"])
tm.assert_frame_equal(result, expected)
def test_variable_width_unicode():
if not compat.PY3:
pytest.skip("Bytes-related test - only needs to work on Python 3")
data = """
שלום שלום
ום שלל
של ום
""".strip("\r\n")
encoding = "utf8"
kwargs = dict(header=None, encoding=encoding)
expected = read_fwf(BytesIO(data.encode(encoding)),
colspecs=[(0, 4), (5, 9)], **kwargs)
result = read_fwf(BytesIO(data.encode(encoding)), **kwargs)
tm.assert_frame_equal(result, expected)
def test_fwf_regression():
# see gh-3594
#
# Turns out "T060" is parsable as a datetime slice!
tz_list = [1, 10, 20, 30, 60, 80, 100]
widths = [16] + [8] * len(tz_list)
names = ["SST"] + ["T%03d" % z for z in tz_list[1:]]
data = """ 2009164202000 9.5403 9.4105 8.6571 7.8372 6.0612 5.8843 5.5192
2009164203000 9.5435 9.2010 8.6167 7.8176 6.0804 5.8728 5.4869
2009164204000 9.5873 9.1326 8.4694 7.5889 6.0422 5.8526 5.4657
2009164205000 9.5810 9.0896 8.4009 7.4652 6.0322 5.8189 5.4379
2009164210000 9.6034 9.0897 8.3822 7.4905 6.0908 5.7904 5.4039
"""
result = read_fwf(StringIO(data), index_col=0, header=None, names=names,
widths=widths, parse_dates=True,
date_parser=lambda s: datetime.strptime(s, "%Y%j%H%M%S"))
expected = DataFrame([
[9.5403, 9.4105, 8.6571, 7.8372, 6.0612, 5.8843, 5.5192],
[9.5435, 9.2010, 8.6167, 7.8176, 6.0804, 5.8728, 5.4869],
[9.5873, 9.1326, 8.4694, 7.5889, 6.0422, 5.8526, 5.4657],
[9.5810, 9.0896, 8.4009, 7.4652, 6.0322, 5.8189, 5.4379],
[9.6034, 9.0897, 8.3822, 7.4905, 6.0908, 5.7904, 5.4039],
], index=DatetimeIndex(["2009-06-13 20:20:00", "2009-06-13 20:30:00",
"2009-06-13 20:40:00", "2009-06-13 20:50:00",
"2009-06-13 21:00:00"]),
columns=["SST", "T010", "T020", "T030", "T060", "T080", "T100"])
tm.assert_frame_equal(result, expected)
def getdept(txt,datereq):
firstheadpat = compile('\d{2}/\d{2}/\d{4}') #not for zip files!
#trainpat = compile('(?<=\* Train )\d+')
lastheadpat = compile('^\* V')
datestr = None
with StringIO(txt) as inpt:
for line in inpt:
tmp = firstheadpat.findall(line)
if len(tmp) >0:
datestr = tmp[0]
if len(lastheadpat.findall(line)) > 0:
if datestr is None:
# must be a zip file where no dates are give
datestr = datereq.strftime('%m/%d/%Y')
break
#data = read_fwf(inpt,colspecs=[(2,5),(10,15),(16,17),(19,24),(25,30),(31,36)],skiprows=0)
#data.columns = ['city','skedarv','skeddep','actarv','actdep']
data = read_fwf(inpt,
colspecs=[(2,5),(16,17),(19,24),(31,36)],
index_col=0,
header=None,
skiprows=0)
#%% append last arrival (destination)
arv = getarv(txt)
#%% drop blank rows before appending arrival
data = data.dropna(axis=0,how='all') #needed for trailing blank lines
data = data.replace('*',nan) #now that blank lines are gone, we swap for nan
data.ix[-1] = arv.ix[0] #we know arrival is one line, the last line of the file
data.columns = ['day','sked','act']
return data, datestr
def test_fwf_compression(compression):
data = """1111111111
2222222222
3333333333""".strip()
kwargs = dict(widths=[5, 5], names=["one", "two"])
expected = read_fwf(StringIO(data), **kwargs)
if compat.PY3:
data = bytes(data, encoding="utf-8")
with tm.ensure_clean() as path:
tm.write_to_compressed(compression, path, data)
result = read_fwf(path, compression=compression, **kwargs)
tm.assert_frame_equal(result, expected)
def test_bytes_io_input():
if not compat.PY3:
pytest.skip("Bytes-related test - only needs to work on Python 3")
result = read_fwf(BytesIO("שלום\nשלום".encode('utf8')),
widths=[2, 2], encoding="utf8")
expected = DataFrame([["של", "ום"]], columns=["של", "ום"])
tm.assert_frame_equal(result, expected)
def test_fwf(self):
data_expected = """\
2011,58,360.242940,149.910199,11950.7
2011,59,444.953632,166.985655,11788.4
2011,60,364.136849,183.628767,11806.2
2011,61,413.836124,184.375703,11916.8
2011,62,502.953953,173.237159,12468.3
"""
expected = read_csv(StringIO(data_expected), header=None)
data1 = """\
201158 360.242940 149.910199 11950.7
201159 444.953632 166.985655 11788.4
201160 364.136849 183.628767 11806.2
201161 413.836124 184.375703 11916.8
201162 502.953953 173.237159 12468.3
"""
colspecs = [(0, 4), (4, 8), (8, 20), (21, 33), (34, 43)]
df = read_fwf(StringIO(data1), colspecs=colspecs, header=None)
assert_frame_equal(df, expected)
data2 = """\
2011 58 360.242940 149.910199 11950.7
2011 59 444.953632 166.985655 11788.4
2011 60 364.136849 183.628767 11806.2
2011 61 413.836124 184.375703 11916.8
2011 62 502.953953 173.237159 12468.3
"""
df = read_fwf(StringIO(data2), widths=[5, 5, 13, 13, 7], header=None)
assert_frame_equal(df, expected)
# From Thomas Kluyver: apparently some non-space filler characters can
# be seen, this is supported by specifying the 'delimiter' character:
# http://publib.boulder.ibm.com/infocenter/dmndhelp/v6r1mx/index.jsp?topic=/com.ibm.wbit.612.help.config.doc/topics/rfixwidth.html
data3 = """\
201158~~~~360.242940~~~149.910199~~~11950.7
201159~~~~444.953632~~~166.985655~~~11788.4
201160~~~~364.136849~~~183.628767~~~11806.2
201161~~~~413.836124~~~184.375703~~~11916.8
201162~~~~502.953953~~~173.237159~~~12468.3
"""
df = read_fwf(StringIO(data3), colspecs=colspecs, delimiter='~', header=None)
assert_frame_equal(df, expected)
self.assertRaises(ValueError, read_fwf, StringIO(data3),
colspecs=colspecs, widths=[6, 10, 10, 7])
def test_BytesIO_input(self):
if not compat.PY3:
raise nose.SkipTest(
"Bytes-related test - only needs to work on Python 3")
result = read_fwf(BytesIO("שלום\nשלום".encode('utf8')), widths=[
2, 2], encoding='utf8')
expected = DataFrame([["של", "ום"]], columns=["של", "ום"])
tm.assert_frame_equal(result, expected)
def test_skiprows_passing_as_positional_deprecated():
# GH#41485
data = """0
1
2
"""
with tm.assert_produces_warning(FutureWarning, match="keyword-only"):
result = read_fwf(StringIO(data), [(0, 2)])
expected = DataFrame({"0": [1, 2]})
tm.assert_frame_equal(result, expected)
def test_1000_fwf(self):
data = """
1 2,334.0 5
10 13 10.
"""
expected = np.array([[1, 2334., 5], [10, 13, 10]])
df = read_fwf(StringIO(data),
colspecs=[(0, 3), (3, 11), (12, 16)],
thousands=',')
tm.assert_almost_equal(df.values, expected)
def test_1000_fwf(self):
data = """
1 2,334.0 5
10 13 10.
"""
expected = [[1, 2334., 5],
[10, 13, 10]]
df = read_fwf(StringIO(data), colspecs=[(0,3),(3,11),(12,16)],
thousands=',')
assert_almost_equal(df.values, expected)
def test_comment_fwf(self):
data = """
1 2. 4 #hello world
5 NaN 10.0
"""
expected = [[1, 2., 4],
[5, np.nan, 10.]]
df = read_fwf(StringIO(data), colspecs=[(0,3),(4,9),(9,25)],
comment='#')
assert_almost_equal(df.values, expected)
def test_comment_fwf(self):
data = """
1 2. 4 #hello world
5 NaN 10.0
"""
expected = np.array([[1, 2., 4], [5, np.nan, 10.]])
df = read_fwf(StringIO(data),
colspecs=[(0, 3), (4, 9), (9, 25)],
comment='#')
tm.assert_almost_equal(df.values, expected)
def test_fwf_skip_blank_lines():
data = """
A B C D
201158 360.242940 149.910199 11950.7
201159 444.953632 166.985655 11788.4
201162 502.953953 173.237159 12468.3
"""
result = read_fwf(StringIO(data), skip_blank_lines=True)
expected = DataFrame(
[
[201158, 360.242940, 149.910199, 11950.7],
[201159, 444.953632, 166.985655, 11788.4],
[201162, 502.953953, 173.237159, 12468.3],
],
columns=["A", "B", "C", "D"],
)
tm.assert_frame_equal(result, expected)
data = """\
A B C D
201158 360.242940 149.910199 11950.7
201159 444.953632 166.985655 11788.4
201162 502.953953 173.237159 12468.3
"""
result = read_fwf(StringIO(data), skip_blank_lines=False)
expected = DataFrame(
[
[201158, 360.242940, 149.910199, 11950.7],
[201159, 444.953632, 166.985655, 11788.4],
[np.nan, np.nan, np.nan, np.nan],
[np.nan, np.nan, np.nan, np.nan],
[201162, 502.953953, 173.237159, 12468.3],
],
columns=["A", "B", "C", "D"],
)
tm.assert_frame_equal(result, expected)
def test_fwf_colspecs_infer_nrows(infer_nrows, exp_data):
# see gh-15138
data = """\
1 2
123 98
"""
expected = DataFrame(exp_data)
result = read_fwf(StringIO(data), infer_nrows=infer_nrows, header=None)
tm.assert_frame_equal(result, expected)
def test_fwf_colspecs_none(colspecs, exp_data):
# see gh-7079
data = """\
123456
456789
"""
expected = DataFrame(exp_data)
result = read_fwf(StringIO(data), colspecs=colspecs, header=None)
tm.assert_frame_equal(result, expected)
def test_BytesIO_input(self):
if not compat.PY3:
raise nose.SkipTest(
"Bytes-related test - only needs to work on Python 3")
result = read_fwf(BytesIO("שלום\nשלום".encode('utf8')),
widths=[2, 2],
encoding='utf8')
expected = DataFrame([["של", "ום"]], columns=["של", "ום"])
tm.assert_frame_equal(result, expected)
def test_skiprows_inference(self):
# GH11256
test = """
Text contained in the file header
DataCol1 DataCol2
0.0 1.0
101.6 956.1
""".strip()
expected = read_csv(StringIO(test), skiprows=2, delim_whitespace=True)
tm.assert_frame_equal(expected, read_fwf(StringIO(test), skiprows=2))
def test_names_and_infer_colspecs():
# GH#45337
data = """X Y Z
959.0 345 22.2
"""
result = read_fwf(StringIO(data),
skiprows=1,
usecols=[0, 2],
names=["a", "b"])
expected = DataFrame({"a": [959.0], "b": 22.2})
tm.assert_frame_equal(result, expected)
def test_fwf_compression(compression_only, infer):
data = """1111111111
2222222222
3333333333""".strip()
compression = compression_only
extension = "gz" if compression == "gzip" else compression
kwargs = dict(widths=[5, 5], names=["one", "two"])
expected = read_fwf(StringIO(data), **kwargs)
data = bytes(data, encoding="utf-8")
with tm.ensure_clean(filename="tmp." + extension) as path:
tm.write_to_compressed(compression, path, data)
if infer is not None:
kwargs["compression"] = "infer" if infer else compression
result = read_fwf(path, **kwargs)
tm.assert_frame_equal(result, expected)
def test_default_delimiter():
header = None
csv_data = """
a,bbb
cc,dd"""
fwf_data = """
a \tbbb
cc\tdd """
result = read_fwf(StringIO(fwf_data), widths=[3, 3], header=header, skiprows=[0])
expected = read_csv(StringIO(csv_data), header=header)
tm.assert_frame_equal(result, expected)
def test_fwf_comment(comment):
data = """\
1 2. 4 #hello world
5 NaN 10.0
"""
data = data.replace("#", comment)
colspecs = [(0, 3), (4, 9), (9, 25)]
expected = DataFrame([[1, 2.0, 4], [5, np.nan, 10.0]])
result = read_fwf(StringIO(data), colspecs=colspecs, header=None, comment=comment)
tm.assert_almost_equal(result, expected)
def test_len_colspecs_len_names_with_index_col(colspecs, names, widths,
index_col, expected):
# GH#40830
data = """col1 col2 col3 col4
bab ba 2"""
result = read_fwf(
StringIO(data),
colspecs=colspecs,
names=names,
widths=widths,
index_col=index_col,
)
tm.assert_frame_equal(result, expected)
def test_fwf(self):
data_expected = """\
2011,58,360.242940,149.910199,11950.7
2011,59,444.953632,166.985655,11788.4
2011,60,364.136849,183.628767,11806.2
2011,61,413.836124,184.375703,11916.8
2011,62,502.953953,173.237159,12468.3
"""
expected = read_csv(StringIO(data_expected),
engine='python',
header=None)
data1 = """\
201158 360.242940 149.910199 11950.7
201159 444.953632 166.985655 11788.4
201160 364.136849 183.628767 11806.2
201161 413.836124 184.375703 11916.8
201162 502.953953 173.237159 12468.3
"""
colspecs = [(0, 4), (4, 8), (8, 20), (21, 33), (34, 43)]
df = read_fwf(StringIO(data1), colspecs=colspecs, header=None)
tm.assert_frame_equal(df, expected)
data2 = """\
2011 58 360.242940 149.910199 11950.7
2011 59 444.953632 166.985655 11788.4
2011 60 364.136849 183.628767 11806.2
2011 61 413.836124 184.375703 11916.8
2011 62 502.953953 173.237159 12468.3
"""
df = read_fwf(StringIO(data2), widths=[5, 5, 13, 13, 7], header=None)
tm.assert_frame_equal(df, expected)
# From Thomas Kluyver: apparently some non-space filler characters can
# be seen, this is supported by specifying the 'delimiter' character:
# http://publib.boulder.ibm.com/infocenter/dmndhelp/v6r1mx/index.jsp?topic=/com.ibm.wbit.612.help.config.doc/topics/rfixwidth.html
data3 = """\
201158~~~~360.242940~~~149.910199~~~11950.7
201159~~~~444.953632~~~166.985655~~~11788.4
201160~~~~364.136849~~~183.628767~~~11806.2
201161~~~~413.836124~~~184.375703~~~11916.8
201162~~~~502.953953~~~173.237159~~~12468.3
"""
df = read_fwf(StringIO(data3),
colspecs=colspecs,
delimiter='~',
header=None)
tm.assert_frame_equal(df, expected)
with tm.assertRaisesRegexp(ValueError, "must specify only one of"):
read_fwf(StringIO(data3), colspecs=colspecs, widths=[6, 10, 10, 7])
with tm.assertRaisesRegexp(ValueError, "Must specify either"):
read_fwf(StringIO(data3), colspecs=None, widths=None)
def test_skiprows_by_index_inference(self):
test = """
To be skipped
Not To Be Skipped
Once more to be skipped
123 34 8 123
456 78 9 456
""".strip()
expected = read_csv(StringIO(test), skiprows=[0, 2],
delim_whitespace=True)
tm.assert_frame_equal(expected, read_fwf(
StringIO(test), skiprows=[0, 2]))
def test_skip_rows_and_n_rows():
# GH#44021
data = """a\tb
1\t a
2\t b
3\t c
4\t d
5\t e
6\t f
"""
result = read_fwf(StringIO(data), nrows=4, skiprows=[2, 4])
expected = DataFrame({"a": [1, 3, 5, 6], "b": ["a", "c", "e", "f"]})
tm.assert_frame_equal(result, expected)
def test_skiprows_by_index_inference():
data = """
To be skipped
Not To Be Skipped
Once more to be skipped
123 34 8 123
456 78 9 456
""".strip()
skiprows = [0, 2]
expected = read_csv(StringIO(data), skiprows=skiprows, delim_whitespace=True)
result = read_fwf(StringIO(data), skiprows=skiprows)
tm.assert_frame_equal(result, expected)
def test_skiprows_inference():
# see gh-11256
data = """
Text contained in the file header
DataCol1 DataCol2
0.0 1.0
101.6 956.1
""".strip()
skiprows = 2
expected = read_csv(StringIO(data), skiprows=skiprows, delim_whitespace=True)
result = read_fwf(StringIO(data), skiprows=skiprows)
tm.assert_frame_equal(result, expected)
def test_fwf_thousands(thousands):
data = """\
1 2,334.0 5
10 13 10.
"""
data = data.replace(",", thousands)
colspecs = [(0, 3), (3, 11), (12, 16)]
expected = DataFrame([[1, 2334.0, 5], [10, 13, 10.0]])
result = read_fwf(
StringIO(data), header=None, colspecs=colspecs, thousands=thousands
)
tm.assert_almost_equal(result, expected)
def test_fwf_compression(self):
try:
import gzip
import bz2
except ImportError:
raise nose.SkipTest("Need gzip and bz2 to run this test")
data = """1111111111
2222222222
3333333333""".strip()
widths = [5, 5]
names = ['one', 'two']
expected = read_fwf(StringIO(data), widths=widths, names=names)
if compat.PY3:
data = bytes(data, encoding='utf-8')
comps = [('gzip', gzip.GzipFile), ('bz2', bz2.BZ2File)]
for comp_name, compresser in comps:
with tm.ensure_clean() as path:
tmp = compresser(path, mode='wb')
tmp.write(data)
tmp.close()
result = read_fwf(path, widths=widths, names=names,
compression=comp_name)
tm.assert_frame_equal(result, expected)
def test_whitespace_preservation():
# see gh-16772
header = None
csv_data = """
a ,bbb
cc,dd """
fwf_data = """
a bbb
ccdd """
result = read_fwf(
StringIO(fwf_data), widths=[3, 3], header=header, skiprows=[0], delimiter="\n\t"
)
expected = read_csv(StringIO(csv_data), header=header)
tm.assert_frame_equal(result, expected)
def test_default_delimiter(self):
data_expected = """
a,bbb
cc,dd"""
expected = read_csv(StringIO(data_expected), header=None)
test_data = """
a \tbbb
cc\tdd """
result = read_fwf(StringIO(test_data),
widths=[3, 3],
header=None,
skiprows=[0])
tm.assert_frame_equal(result, expected)
def test_dtype(dtype):
data = """ a b c
1 2 3.2
3 4 5.2
"""
colspecs = [(0, 5), (5, 10), (10, None)]
result = read_fwf(StringIO(data), colspecs=colspecs, dtype=dtype)
expected = pd.DataFrame(
{"a": [1, 3], "b": [2, 4], "c": [3.2, 5.2]}, columns=["a", "b", "c"]
)
for col, dt in dtype.items():
expected[col] = expected[col].astype(dt)
tm.assert_frame_equal(result, expected)
def test_basic():
data = """\
A B C D
201158 360.242940 149.910199 11950.7
201159 444.953632 166.985655 11788.4
201160 364.136849 183.628767 11806.2
201161 413.836124 184.375703 11916.8
201162 502.953953 173.237159 12468.3
"""
result = read_fwf(StringIO(data))
expected = DataFrame([[201158, 360.242940, 149.910199, 11950.7],
[201159, 444.953632, 166.985655, 11788.4],
[201160, 364.136849, 183.628767, 11806.2],
[201161, 413.836124, 184.375703, 11916.8],
[201162, 502.953953, 173.237159, 12468.3]],
columns=["A", "B", "C", "D"])
tm.assert_frame_equal(result, expected)
def test_whitespace_preservation(self):
# Addresses Issue #16772
data_expected = """
a ,bbb
cc,dd """
expected = read_csv(StringIO(data_expected), header=None)
test_data = """
a bbb
ccdd """
result = read_fwf(StringIO(test_data),
widths=[3, 3],
header=None,
skiprows=[0],
delimiter="\n\t")
tm.assert_frame_equal(result, expected)
def test_binary_mode():
"""
read_fwf supports opening files in binary mode.
GH 18035.
"""
data = """aas aas aas
bba bab b a"""
df_reference = DataFrame(
[["bba", "bab", "b a"]], columns=["aas", "aas.1", "aas.2"], index=[0]
)
with tm.ensure_clean() as path:
Path(path).write_text(data)
with open(path, "rb") as file:
df = read_fwf(file)
file.seek(0)
tm.assert_frame_equal(df, df_reference)
def test_encoding_mmap(memory_map):
"""
encoding should be working, even when using a memory-mapped file.
GH 23254.
"""
encoding = "iso8859_1"
with tm.ensure_clean() as path:
Path(path).write_bytes(" 1 A Ä 2\n".encode(encoding))
df = read_fwf(
path,
header=None,
widths=[2, 2, 2, 2],
encoding=encoding,
memory_map=memory_map,
)
df_reference = DataFrame([[1, "A", "Ä", 2]])
tm.assert_frame_equal(df, df_reference)
def test_encoding_mmap(memory_map):
"""
encoding should be working, even when using a memory-mapped file.
GH 23254.
"""
encoding = "iso8859_1"
data = BytesIO(" 1 A Ä 2\n".encode(encoding))
df = read_fwf(
data,
header=None,
widths=[2, 2, 2, 2],
encoding=encoding,
memory_map=memory_map,
)
data.seek(0)
df_reference = DataFrame([[1, "A", "Ä", 2]])
tm.assert_frame_equal(df, df_reference)
def test_fwf_regression():
# see gh-3594
#
# Turns out "T060" is parsable as a datetime slice!
tz_list = [1, 10, 20, 30, 60, 80, 100]
widths = [16] + [8] * len(tz_list)
names = ["SST"] + [f"T{z:03d}" for z in tz_list[1:]]
data = """ 2009164202000 9.5403 9.4105 8.6571 7.8372 6.0612 5.8843 5.5192
2009164203000 9.5435 9.2010 8.6167 7.8176 6.0804 5.8728 5.4869
2009164204000 9.5873 9.1326 8.4694 7.5889 6.0422 5.8526 5.4657
2009164205000 9.5810 9.0896 8.4009 7.4652 6.0322 5.8189 5.4379
2009164210000 9.6034 9.0897 8.3822 7.4905 6.0908 5.7904 5.4039
"""
result = read_fwf(
StringIO(data),
index_col=0,
header=None,
names=names,
widths=widths,
parse_dates=True,
date_parser=lambda s: datetime.strptime(s, "%Y%j%H%M%S"),
)
expected = DataFrame(
[
[9.5403, 9.4105, 8.6571, 7.8372, 6.0612, 5.8843, 5.5192],
[9.5435, 9.2010, 8.6167, 7.8176, 6.0804, 5.8728, 5.4869],
[9.5873, 9.1326, 8.4694, 7.5889, 6.0422, 5.8526, 5.4657],
[9.5810, 9.0896, 8.4009, 7.4652, 6.0322, 5.8189, 5.4379],
[9.6034, 9.0897, 8.3822, 7.4905, 6.0908, 5.7904, 5.4039],
],
index=DatetimeIndex(
[
"2009-06-13 20:20:00",
"2009-06-13 20:30:00",
"2009-06-13 20:40:00",
"2009-06-13 20:50:00",
"2009-06-13 21:00:00",
]
),
columns=["SST", "T010", "T020", "T030", "T060", "T080", "T100"],
)
tm.assert_frame_equal(result, expected)
def test_widths_and_usecols():
# GH#46580
data = """0 1 n -0.4100.1
0 2 p 0.2 90.1
0 3 n -0.3140.4"""
result = read_fwf(
StringIO(data),
header=None,
usecols=(0, 1, 3),
widths=(3, 5, 1, 5, 5),
index_col=False,
names=("c0", "c1", "c3"),
)
expected = DataFrame({
"c0": 0,
"c1": [1, 2, 3],
"c3": [-0.4, 0.2, -0.3],
})
tm.assert_frame_equal(result, expected)
