Merge branch 'master' of https://github.com/pandas-dev/pandas into fuse_kh

Author: jbrockmendel

pandas/_libs/hashtable_class_helper.pxi.in

@@ -925,7 +925,8 @@ cdef class PyObjectHashTable(HashTable):
         return self.table.size
 
     def __contains__(self, object key):
-        cdef khiter_t k
+        cdef:
+            khiter_t k
         hash(key)
 
         k = kh_get_pymap(self.table, <PyObject*>key)
@@ -938,7 +939,8 @@ cdef class PyObjectHashTable(HashTable):
                                        sizeof(uint32_t))  # flags
 
     cpdef get_item(self, object val):
-        cdef khiter_t k
+        cdef:
+            khiter_t k
 
         k = kh_get_pymap(self.table, <PyObject*>val)
         if k != self.table.n_buckets:

pandas/_libs/hashtable_func_helper.pxi.in

@@ -82,7 +82,7 @@ cdef inline khiter_t kh_get(table_t *table, scalar_t val) nogil:
 
 {{py:
 
-# dtype, ttype
+# dtype, ttype, c_type
 dtypes = [('float64', 'float64', 'float64_t'),
           ('uint64', 'uint64', 'uint64_t'),
           ('object', 'pymap', 'object'),
@@ -284,7 +284,6 @@ def ismember_{{dtype}}(ndarray[{{c_type}}] arr, ndarray[{{c_type}}] values):
 {{else}}
 def ismember_{{dtype}}({{c_type}}[:] arr, {{c_type}}[:] values):
 {{endif}}
-
     """
     Return boolean of values in arr on an
     element by-element basis

pandas/_libs/internals.pyx

@@ -284,7 +284,7 @@ cdef slice_get_indices_ex(slice slc, Py_ssize_t objlen=PY_SSIZE_T_MAX):
     return start, stop, step, length
 
 
-def slice_getitem(slice slc not None, ind):
+cdef slice_getitem(slice slc, ind):
     cdef:
         Py_ssize_t s_start, s_stop, s_step, s_len
         Py_ssize_t ind_start, ind_stop, ind_step, ind_len

pandas/_libs/interval.pyx

@@ -18,7 +18,7 @@ cnp.import_array()
 
 cimport pandas._libs.util as util
 
-from pandas._libs.hashtable cimport Int64Vector, Int64VectorData
+from pandas._libs.hashtable cimport Int64Vector
 from pandas._libs.tslibs.util cimport is_integer_object, is_float_object
 
 from pandas._libs.tslibs import Timestamp

pandas/_libs/lib.pyx

@@ -9,12 +9,9 @@ import warnings
 import cython
 from cython import Py_ssize_t
 
-from cpython.list cimport PyList_New
-from cpython.object cimport (PyObject_Str, PyObject_RichCompareBool, Py_EQ,
-                             Py_SIZE)
+from cpython.object cimport PyObject_RichCompareBool, Py_EQ
 from cpython.ref cimport Py_INCREF
 from cpython.tuple cimport PyTuple_SET_ITEM, PyTuple_New
-from cpython.unicode cimport PyUnicode_Join
 
 from cpython.datetime cimport (PyDateTime_Check, PyDate_Check,
                                PyTime_Check, PyDelta_Check,

pandas/core/dtypes/cast.py

@@ -202,7 +202,7 @@ def trans(x):
         r = result.ravel()
         arr = np.array([r[0]])
 
-        if isna(arr).any() or not np.allclose(arr, trans(arr).astype(dtype), rtol=0):
+        if isna(arr).any():
             # if we have any nulls, then we are done
             return result
 

pandas/core/generic.py

@@ -10852,7 +10852,7 @@ def transform(self, func, *args, **kwargs):
         Also returns None for empty %(klass)s.
         """
 
-    def _find_valid_index(self, how):
+    def _find_valid_index(self, how: str):
         """
         Retrieves the index of the first valid value.
 

pandas/core/groupby/generic.py

@@ -262,7 +262,7 @@ def aggregate(self, func=None, *args, **kwargs):
 
             try:
                 return self._python_agg_general(func, *args, **kwargs)
-            except AssertionError:
+            except (AssertionError, TypeError):
                 raise
             except Exception:
                 result = self._aggregate_named(func, *args, **kwargs)
@@ -1099,10 +1099,7 @@ def _aggregate_item_by_item(self, func, *args, **kwargs):
 
             cast = self._transform_should_cast(func)
             try:
-
                 result[item] = colg.aggregate(func, *args, **kwargs)
-                if cast:
-                    result[item] = self._try_cast(result[item], data)
 
             except ValueError as err:
                 if "Must produce aggregated value" in str(err):
@@ -1111,10 +1108,10 @@ def _aggregate_item_by_item(self, func, *args, **kwargs):
                     raise
                 cannot_agg.append(item)
                 continue
-            except TypeError as e:
-                cannot_agg.append(item)
-                errors = e
-                continue
+
+            else:
+                if cast:
+                    result[item] = self._try_cast(result[item], data)
 
         result_columns = obj.columns
         if cannot_agg:

pandas/core/util/hashing.py

@@ -5,8 +5,8 @@
 
 import numpy as np
 
+from pandas._libs import Timestamp
 import pandas._libs.hashing as hashing
-import pandas._libs.tslibs as tslibs
 
 from pandas.core.dtypes.cast import infer_dtype_from_scalar
 from pandas.core.dtypes.common import (
@@ -337,8 +337,8 @@ def _hash_scalar(val, encoding: str = "utf8", hash_key=None):
         # for tz-aware datetimes, we need the underlying naive UTC value and
         # not the tz aware object or pd extension type (as
         # infer_dtype_from_scalar would do)
-        if not isinstance(val, tslibs.Timestamp):
-            val = tslibs.Timestamp(val)
+        if not isinstance(val, Timestamp):
+            val = Timestamp(val)
         val = val.tz_convert(None)
 
     dtype, val = infer_dtype_from_scalar(val)

pandas/tests/dtypes/cast/test_downcast.py

@@ -1,3 +1,5 @@
+import decimal
+
 import numpy as np
 import pytest
 
@@ -25,6 +27,13 @@
             "infer",
             np.array([8, 8, 8, 8, 9], dtype=np.int64),
         ),
+        (
+            # This is a judgement call, but we do _not_ downcast Decimal
+            #  objects
+            np.array([decimal.Decimal(0.0)]),
+            "int64",
+            np.array([decimal.Decimal(0.0)]),
+        ),
     ],
 )
 def test_downcast(arr, expected, dtype):

pandas/tests/series/test_repr.py

@@ -17,10 +17,8 @@
 from pandas.core.index import MultiIndex
 import pandas.util.testing as tm
 
-from .common import TestData
 
-
-class TestSeriesRepr(TestData):
+class TestSeriesRepr:
     def test_multilevel_name_print(self):
         index = MultiIndex(
             levels=[["foo", "bar", "baz", "qux"], ["one", "two", "three"]],
@@ -67,24 +65,24 @@ def test_name_printing(self):
         s = Series(index=date_range("20010101", "20020101"), name="test")
         assert "Name: test" in repr(s)
 
-    def test_repr(self):
-        str(self.ts)
-        str(self.series)
-        str(self.series.astype(int))
-        str(self.objSeries)
+    def test_repr(self, datetime_series, string_series, object_series):
+        str(datetime_series)
+        str(string_series)
+        str(string_series.astype(int))
+        str(object_series)
 
         str(Series(tm.randn(1000), index=np.arange(1000)))
         str(Series(tm.randn(1000), index=np.arange(1000, 0, step=-1)))
 
         # empty
-        str(self.empty)
+        str(Series())
 
         # with NaNs
-        self.series[5:7] = np.NaN
-        str(self.series)
+        string_series[5:7] = np.NaN
+        str(string_series)
 
         # with Nones
-        ots = self.ts.astype("O")
+        ots = datetime_series.astype("O")
         ots[::2] = None
         repr(ots)
 
@@ -102,8 +100,8 @@ def test_repr(self):
             ("\u03B1", "\u03B2", "\u03B3"),
             ("\u03B1", "bar"),
         ]:
-            self.series.name = name
-            repr(self.series)
+            string_series.name = name
+            repr(string_series)
 
         biggie = Series(
             tm.randn(1000), index=np.arange(1000), name=("foo", "bar", "baz")