def test_temp_ssh_files():
with filetext('name,balance\nAlice,100\nBob,200', extension='csv') as fn:
csv = CSV(fn)
scsv = into(Temp(SSH(CSV)), csv, hostname='localhost')
assert discover(csv) == discover(scsv)
assert isinstance(scsv, _Temp)
def test_pandas_read_supports_read_csv_kwargs():
with filetext("Alice,1\nBob,2") as fn:
ds = datashape.dshape("var * {name: string, amount: int}")
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds, usecols=["name"])
assert isinstance(df, pd.DataFrame)
assert convert(list, df) == [("Alice",), ("Bob",)]
def test_header_mix_str_digits():
ds = datashape.dshape('''var * {"On- or Off- Budget": ?string,
"1990": ?string}''')
with filetext('On- or Off- Budget,1990\nOn Budget,-628\nOff budget,"5,962"\n') as fn:
csv = CSV(fn, has_header=True)
df = convert(pd.DataFrame, csv)
assert discover(csv).measure == ds.measure
def test_more_unicode_column_names():
with filetext(b'foo\xc4\x87,a\n1,2\n3,4', extension='csv',
mode='wb') as fn:
df = into(pd.DataFrame, CSV(fn, has_header=True))
expected = pd.DataFrame([(1, 2), (3, 4)],
columns=[b'foo\xc4\x87'.decode('utf8'), u'a'])
tm.assert_frame_equal(df, expected)
def test_csv_with_header():
with tmpfile('db') as dbfilename:
with filetext('a,b\n1,2\n3,4', extension='csv') as csvfilename:
t = into('sqlite:///%s::mytable' % dbfilename,
csvfilename, has_header=True)
assert discover(t) == dshape('var * {a: int64, b: int64}')
assert into(set, t) == set([(1, 2), (3, 4)])
def test_pandas_read_supports_read_csv_kwargs():
with filetext('Alice,1\nBob,2') as fn:
ds = datashape.dshape('var * {name: string, amount: int}')
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds, usecols=['name'])
assert isinstance(df, pd.DataFrame)
assert convert(list, df) == [('Alice',), ('Bob',)]
def test_pandas_read():
with filetext('Alice,1\nBob,2') as fn:
ds = datashape.dshape('var * {name: string, amount: int}')
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds)
assert isinstance(df, pd.DataFrame)
assert convert(list, df) == [('Alice', 1), ('Bob', 2)]
assert list(df.columns) == ['name', 'amount']
def test_pandas_read_supports_gzip():
with filetext("Alice,1\nBob,2", open=gzip.open, mode="wt", extension=".csv.gz") as fn:
ds = datashape.dshape("var * {name: string, amount: int}")
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds)
assert isinstance(df, pd.DataFrame)
assert convert(list, df) == [("Alice", 1), ("Bob", 2)]
assert list(df.columns) == ["name", "amount"]
def test_pandas_read_supports_missing_integers():
with filetext("Alice,1\nBob,") as fn:
ds = datashape.dshape("var * {name: string, val: ?int32}")
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds)
assert isinstance(df, pd.DataFrame)
assert list(df.columns) == ["name", "val"]
assert df.dtypes["val"] == "f4"
def test_pandas_read():
with filetext("Alice,1\nBob,2") as fn:
ds = datashape.dshape("var * {name: string, amount: int}")
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds)
assert isinstance(df, pd.DataFrame)
assert convert(list, df) == [("Alice", 1), ("Bob", 2)]
assert list(df.columns) == ["name", "amount"]
def test_header_disagrees_with_dshape():
ds = datashape.dshape("var * {name: string, bal: int64}")
with filetext("name,val\nAlice,100\nBob,200", extension="csv") as fn:
csv = CSV(fn, header=True)
assert convert(list, csv) == [("Alice", 100), ("Bob", 200)]
assert list(convert(pd.DataFrame, csv).columns) == ["name", "val"]
assert list(convert(pd.DataFrame, csv, dshape=ds).columns) == ["name", "bal"]
def test_pandas_read_supports_datetimes():
with filetext('Alice,2014-01-02\nBob,2014-01-03') as fn:
ds = datashape.dshape('var * {name: string, when: date}')
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds)
assert isinstance(df, pd.DataFrame)
assert list(df.columns) == ['name', 'when']
assert df.dtypes['when'] == 'M8[ns]'
def test_csv_to_bcolz():
with filetext('name,runway,takeoff,datetime_nearest_close\n'
'S28,28,TRUE,A\n'
'S16,16,TRUE,Q\n'
'L14,14,FALSE,I', extension='csv') as src:
with tmpfile('bcolz') as tgt:
bc = into(tgt, src)
assert len(bc) == 3
def test_pandas_read_supports_missing_integers():
with filetext('Alice,1\nBob,') as fn:
ds = datashape.dshape('var * {name: string, val: ?int32}')
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds)
assert isinstance(df, pd.DataFrame)
assert list(df.columns) == ['name', 'val']
assert df.dtypes['val'] == 'f4'
def test_copy_local_files_to_hdfs():
with tmpfile_hdfs() as target:
with filetext('name,amount\nAlice,100\nBob,200') as source:
csv = CSV(source)
scsv = HDFS(CSV)(target, hdfs=hdfs)
into(scsv, csv, blocksize=10) # 10 bytes per message
assert discover(scsv) == discover(csv)
def test_header_disagrees_with_dshape():
ds = datashape.dshape('var * {name: string, bal: int64}')
with filetext('name,val\nAlice,100\nBob,200', extension='csv') as fn:
csv = CSV(fn, header=True)
assert convert(list, csv) == [('Alice', 100), ('Bob', 200)]
assert list(convert(pd.DataFrame, csv).columns) == ['name', 'val']
assert list(convert(pd.DataFrame, csv, dshape=ds).columns) == ['name', 'bal']
def test_pandas_read_supports_datetimes():
with filetext("Alice,2014-01-02\nBob,2014-01-03") as fn:
ds = datashape.dshape("var * {name: string, when: date}")
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds)
assert isinstance(df, pd.DataFrame)
assert list(df.columns) == ["name", "when"]
assert df.dtypes["when"] == "M8[ns]"
def test_pandas_loads_in_datetimes_naively():
with filetext('name,when\nAlice,2014-01-01\nBob,2014-02-02') as fn:
csv = CSV(fn, has_header=True)
ds = datashape.dshape('var * {name: ?string, when: ?datetime}')
assert discover(csv) == ds
df = convert(pd.DataFrame, csv)
assert df.dtypes['when'] == 'M8[ns]'
def test_pandas_read_supports_gzip():
with filetext('Alice,1\nBob,2', open=gzip.open,
mode='wt', extension='.csv.gz') as fn:
ds = datashape.dshape('var * {name: string, amount: int}')
csv = CSV(fn)
df = csv_to_dataframe(csv, dshape=ds)
assert isinstance(df, pd.DataFrame)
assert convert(list, df) == [('Alice', 1), ('Bob', 2)]
assert list(df.columns) == ['name', 'amount']
def test_into_double_string(f):
with filetext('alice,1\nbob,2', extension='.csv') as source:
assert odo(source, list) == [('alice', 1), ('bob', 2)]
with tmpfile('.csv') as target:
csv = odo(source, f(target))
assert isinstance(csv, CSV)
with open(target, 'rU') as f:
assert 'alice' in f.read()
def test_ssh_csv_to_s3_csv():
# for some reason this can only be run in the same file as other ssh tests
# and must be a Temp(SSH(CSV)) otherwise tests above this one fail
s3_bucket = pytest.importorskip('odo.backends.tests.test_aws').s3_bucket
with filetext('name,balance\nAlice,100\nBob,200', extension='csv') as fn:
remote = into(Temp(SSH(CSV)), CSV(fn), hostname='localhost')
with s3_bucket('.csv') as b:
result = into(b, remote)
assert discover(result) == discover(resource(b))
def test_hive_creation_from_local_file():
with filetext(accounts_1_csv, extension='csv') as fn:
with hive_table(host) as uri:
t = into(uri, fn, **auth)
assert isinstance(t, sa.Table)
assert into(set, t) == into(set, fn)
t2 = into(uri, fn, **auth)
assert isinstance(t2, sa.Table)
assert len(into(list, t2)) == 2 * len(into(list, fn))
def test_sqlite_to_csv(sep, header):
with tmpfile("db") as dbfilename:
with filetext("a,b\n1,2\n3,4", extension="csv") as csvfilename:
t = odo(csvfilename, "sqlite:///%s::mytable" % dbfilename)
with tmpfile(".csv") as fn:
odo(t, fn, header=header, delimiter=sep)
with open(fn, "rt") as f:
lines = f.readlines()
expected = [tuple(map(int, row)) for row in map(lambda x: x.split(sep), lines[header:])]
assert odo(fn, list, delimiter=sep, has_header=header, dshape=discover(t)) == expected
def test_different_encoding_to_csv():
with tmpfile('db') as dbfilename:
with filetext('a,b\n1,2\n3,4', extension='csv') as csvfilename:
t = odo(
csvfilename,
'sqlite:///%s::mytable' % dbfilename,
encoding='latin1'
)
with tmpfile('.csv') as fn:
with pytest.raises(ValueError):
odo(t, fn, encoding='latin1')
def test_send_parameterized_query_to_csv():
with tmpfile('db') as dbfilename:
with filetext('a,b\n1,2\n3,4', extension='csv') as csvfilename:
t = odo(
csvfilename,
'sqlite:///%s::mytable' % dbfilename,
)
with tmpfile('.csv') as fn:
q = t.select(t.c.a == 1)
r = odo(q, fn)
assert sorted(odo(q, list)) == sorted(odo(r, list))
def accounts_ssh():
""" Three csv files on the remote host in a directory """
dirname = str(uuid.uuid1())
conn = sftp(**auth)
conn.mkdir(dirname)
with filetext(accounts_1_csv) as fn:
conn.put(fn, dirname + '/accounts.1.csv')
with filetext(accounts_2_csv) as fn:
conn.put(fn, dirname + '/accounts.2.csv')
with filetext(accounts_3_csv) as fn:
conn.put(fn, dirname + '/accounts.3.csv')
filenames = [dirname + '/accounts.%d.csv' % i for i in [1, 2, 3]]
uris = ['ssh://ubuntu@%s:%s' % (host, fn) for fn in filenames]
try:
yield 'ssh://ubuntu@%s:%s/*.csv' % (host, dirname), uris
finally:
for fn in filenames:
conn.remove(fn)
conn.rmdir(dirname)
def test_convert_through_temporary_local_storage():
with filetext('name,quantity\nAlice,100\nBob,200', extension='csv') as fn:
csv = CSV(fn)
df = into(pd.DataFrame, csv)
scsv = into(Temp(SSH(CSV)), csv, hostname='localhost')
assert into(list, csv) == into(list, scsv)
scsv2 = into(Temp(SSH(CSV)), df, hostname='localhost')
assert into(list, scsv2) == into(list, df)
sjson = into(Temp(SSH(JSONLines)), df, hostname='localhost')
assert (into(np.ndarray, sjson) == into(np.ndarray, df)).all()
def test_dialect_of():
with filetext(accounts_1_csv) as fn:
d = dialect_of(CSV(fn))
assert d['delimiter'] == ','
assert d['has_header'] is True
with accounts_data() as (directory, (a, b, c)):
directory2 = HDFS(Directory(CSV))(directory.path, hdfs=directory.hdfs)
d = dialect_of(directory2)
assert d['has_header'] is True
directory2 = HDFS(Directory(CSV))(directory.path, hdfs=directory.hdfs,
has_header=False)
d = dialect_of(directory2)
assert d['has_header'] is False
def test_copy_remote_csv():
with tmpfile('csv') as target:
with filetext('name,balance\nAlice,100\nBob,200',
extension='csv') as fn:
csv = resource(fn)
uri = 'ssh://localhost:%s.csv' % target
scsv = into(uri, csv)
assert isinstance(scsv, SSH(CSV))
assert discover(scsv) == discover(csv)
# Round trip
csv2 = into(target, scsv)
assert into(list, csv) == into(list, csv2)
def test_drop():
with filetext('name,balance\nAlice,100\nBob,200', extension='csv') as fn:
with tmpfile('csv') as target:
scsv = SSH(CSV)(target, hostname='localhost')
assert not os.path.exists(target)
conn = sftp(**scsv.auth)
conn.put(fn, target)
assert os.path.exists(target)
drop(scsv)
drop(scsv)
assert not os.path.exists(target)
def test_sqlite_to_csv(sep, header):
with tmpfile('db') as dbfilename:
with filetext('a,b\n1,2\n3,4', extension='csv') as csvfilename:
t = odo(csvfilename, 'sqlite:///%s::mytable' % dbfilename)
with tmpfile('.csv') as fn:
odo(t, fn, header=header, delimiter=sep)
with open(fn, 'rt') as f:
lines = f.readlines()
expected = [
tuple(map(int, row))
for row in map(lambda x: x.split(sep), lines[header:])
]
assert odo(fn,
list,
delimiter=sep,
has_header=header,
dshape=discover(t)) == expected
def test_unicode_column_names():
with filetext(b'f\xc3\xbc,a\n1,2\n3,4', extension='csv', mode='wb') as fn:
df = into(pd.DataFrame, CSV(fn, has_header=True))
expected = pd.DataFrame([(1, 2), (3, 4)],
columns=[b'f\xc3\xbc'.decode('utf8'), u'a'])
tm.assert_frame_equal(df, expected)
def test_csv_missing_values():
with filetext('name,val\nAlice,100\nNA,200', extension='csv') as fn:
csv = CSV(fn)
assert discover(csv).measure.dict['name'] == Option(string)
def test_csv_separator_header():
with filetext('a|b|c\n1|2|3\n4|5|6', extension='csv') as fn:
csv = CSV(fn, delimiter='|', has_header=True)
assert convert(list, csv) == [(1, 2, 3), (4, 5, 6)]
def test_into_string_on_right():
with filetext('alice,1\nbob,2', extension='.csv') as source:
assert into([], source) == [('alice', 1), ('bob', 2)]
def test_into_string_on_right(f):
with filetext('alice,1\nbob,2', extension='.csv') as source:
assert odo(f(source), []) == [('alice', 1), ('bob', 2)]
def test_unicode_column_names():
with filetext('foo\xc4\x87,a\n1,2\n3,4', extension='csv') as fn:
csv = CSV(fn, has_header=True)
df = into(pd.DataFrame, csv)
def test_pandas_read_supports_whitespace_strings():
with filetext('a,b, \n1,2, \n2,3, \n', extension='csv') as fn:
csv = CSV(fn)
ds = discover(csv)
assert ds == datashape.dshape("var * {a: int64, b: int64, '': ?string}")
def test_header_argument_set_with_or_without_header():
with filetext('name,val\nAlice,100\nBob,200', extension='csv') as fn:
assert into(list, fn) == [('Alice', 100), ('Bob', 200)]
with filetext('Alice,100\nBob,200', extension='csv') as fn:
assert into(list, fn) == [('Alice', 100), ('Bob', 200)]
def test_discover():
with filetext('name,balance\nAlice,100\nBob,200') as fn:
local = CSV(fn)
remote = SSH(CSV)(fn, hostname='localhost')
assert discover(local) == discover(remote)
def test_discover_csv_yields_string_on_totally_empty_columns():
expected = dshape('var * {a: int64, b: ?string, c: int64}')
with filetext('a,b,c\n1,,3\n4,,6\n7,,9') as fn:
csv = CSV(fn, has_header=True)
assert discover(csv) == expected
def test_discover_csv_without_columns():
with filetext('Alice,100\nBob,200', extension='csv') as fn:
csv = CSV(fn)
ds = discover(csv)
assert '100' not in str(ds)
def test_discover_csv_files_without_header():
with filetext('Alice,2014-01-01\nBob,2014-02-02') as fn:
csv = CSV(fn, has_header=False)
df = convert(pd.DataFrame, csv)
assert len(df) == 2
assert 'Alice' not in list(df.columns)
def test_csv_into_list():
with filetext('name,val\nAlice,100\nBob,200', extension='csv') as fn:
L = into(list, fn)
assert L == [('Alice', 100), ('Bob', 200)]
def test_pandas_discover_on_gzipped_files():
with filetext('name,when\nAlice,2014-01-01\nBob,2014-02-02',
open=gzip.open, mode='wt', extension='.csv.gz') as fn:
csv = CSV(fn, has_header=True)
ds = datashape.dshape('var * {name: ?string, when: ?datetime}')
assert discover(csv) == ds
def test_convert():
with filetext('Hello\nWorld') as fn:
assert convert(list, TextFile(fn)) == ['Hello\n', 'World']
def test_drop():
with filetext('hello\nworld') as fn:
t = TextFile(fn)
assert os.path.exists(fn)
drop(t)
assert not os.path.exists(fn)
def test_infer_header():
with filetext('name,val\nAlice,100\nNA,200', extension='csv') as fn:
assert infer_header(CSV(fn).path, 100) == True
with filetext('Alice,100\nNA,200', extension='csv') as fn:
assert infer_header(CSV(fn).path, 100) == False
def test_raise_errors_quickly_on_into_chunks_dataframe():
with filetext('name,val\nAlice,100\nBob,foo', extension='csv') as fn:
ds = datashape.dshape('var * {name: string, val: int}')
csv = CSV(fn, header=True)
assert raises(Exception,
lambda: CSV_to_chunks_of_dataframes(csv, dshape=ds))
def test_csv_infer_header():
with tmpfile('db') as dbfilename:
with filetext('a,b\n1,2\n3,4', extension='csv') as csvfilename:
t = odo(csvfilename, 'sqlite:///%s::mytable' % dbfilename)
assert discover(t) == dshape('var * {a: int64, b: int64}')
assert odo(t, set) == set([(1, 2), (3, 4)])
def test_unused_datetime_columns():
ds = datashape.dshape('var * {val: string, when: datetime}')
with filetext("val,when\na,2000-01-01\nb,2000-02-02") as fn:
csv = CSV(fn, has_header=True)
assert convert(list, csv_to_dataframe(csv, usecols=['val'],
squeeze=True, dshape=ds)) == ['a', 'b']
def test_convert_local_file_to_temp_ssh_file():
with filetext('name,balance\nAlice,100\nBob,200', extension='csv') as fn:
csv = CSV(fn)
scsv = convert(Temp(SSH(CSV)), csv, hostname='localhost')
assert into(list, csv) == into(list, scsv)
def test_empty_dataframe():
with filetext('name,val', extension='csv') as fn:
csv = CSV(fn, has_header=True)
df = convert(pd.DataFrame, csv)
assert isinstance(df, pd.DataFrame)
def test_discover_csv_with_spaces_in_header():
with filetext(' name, val\nAlice,100\nBob,200', extension='csv') as fn:
ds = discover(CSV(fn, has_header=True))
assert ds.measure.names == ['name', 'val']
def test_discover_from_resource():
with filetext('name,balance\nAlice,100\nBob,200', extension='csv') as fn:
local = CSV(fn)
remote = resource('ssh://localhost:' + fn)
assert discover(local) == discover(remote)
