MAINT: Make take_1d accept readonly buffers., #11052

This is a straightforward port of GH#10070 to 1d arrays.

Author: Joris Vankerschaver

doc/source/whatsnew/v0.18.0.txt

@@ -440,10 +440,7 @@ Bug Fixes
 - Bug in ``date_range`` when the boundaries fell on the frequency (:issue:`11804`)
 - Bug in consistency of passing nested dicts to ``.groupby(...).agg(...)`` (:issue:`9052`)
 - Accept unicode in ``Timedelta`` constructor (:issue:`11995`)
-
 - Bug in value label reading for ``StataReader`` when reading incrementally (:issue:`12014`)
-
-
 - Bug in vectorized ``DateOffset`` when ``n`` parameter is ``0`` (:issue:`11370`)
 
 
@@ -456,7 +453,6 @@ Bug Fixes
 - Bug in ``.loc`` against ``CategoricalIndex`` may result in normal ``Index`` (:issue:`11586`)
 - Bug in ``DataFrame.info`` when duplicated column names exist (:issue:`11761`)
 - Bug in ``.copy`` of datetime tz-aware objects (:issue:`11794`)
-
 - Bug in ``Series.apply`` and ``Series.map`` where ``timedelta64`` was not boxed (:issue:`11349`)
 
 
@@ -465,48 +461,39 @@ Bug Fixes
 - Bug in subclasses of ``DataFrame`` where ``AttributeError`` did not propagate (:issue:`11808`)
 - Bug groupby on tz-aware data where selection not returning ``Timestamp`` (:issue:`11616`)
 - Bug in ``pd.read_clipboard`` and ``pd.to_clipboard`` functions not supporting Unicode; upgrade included ``pyperclip`` to v1.5.15 (:issue:`9263`)
-
-
 - Bug in ``DataFrame.query`` containing an assignment (:issue:`8664`)
 
 
 
 - Bug in timezone info lost when broadcasting scalar datetime to ``DataFrame`` (:issue:`11682`)
-
 - Bug in ``Index`` creation from ``Timestamp`` with mixed tz coerces to UTC (:issue:`11488`)
 - Bug in ``to_numeric`` where it does not raise if input is more than one dimension (:issue:`11776`)
-
 - Bug in parsing timezone offset strings with non-zero minutes (:issue:`11708`)
-
 - Bug in ``df.plot`` using incorrect colors for bar plots under matplotlib 1.5+ (:issue:`11614`)
 - Bug in the ``groupby`` ``plot`` method when using keyword arguments (:issue:`11805`).
-
 - Bug in ``DataFrame.duplicated`` and ``drop_duplicates`` causing spurious matches when setting ``keep=False`` (:issue:`11864`)
-
 - Bug in ``.loc`` result with duplicated key may have ``Index`` with incorrect dtype (:issue:`11497`)
 - Bug in ``pd.rolling_median`` where memory allocation failed even with sufficient memory (:issue:`11696`)
-
 - Bug in ``.style.bar`` may not rendered properly using specific browser (:issue:`11678`)
-
 - Bug in rich comparison of ``Timedelta`` with a ``numpy.array`` of ``Timedelta`` that caused an infinite recursion (:issue:`11835`)
-
 - Bug in ``DataFrame.round`` dropping column index name (:issue:`11986`)
-
 - Bug in ``df.replace`` while replacing value in mixed dtype ``Dataframe`` (:issue:`11698`)
-
 - Bug in ``Index`` prevents copying name of passed ``Index``, when a new name is not provided (:issue:`11193`)
-
 - Bug in ``read_excel`` failing to read any non-empty sheets when empty sheets exist and ``sheetname=None`` (:issue:`11711`)
-
 - Bug in ``read_excel`` failing to raise ``NotImplemented`` error when keywords ``parse_dates`` and ``date_parser`` are provided (:issue:`11544`)
-
 - Bug in ``read_sql`` with pymysql connections failing to return chunked data (:issue:`11522`)
-
 - Bug in ``.to_csv`` ignoring formatting parameters ``decimal``, ``na_rep``, ``float_format`` for float indexes (:issue:`11553`)
-
 - Bug in ``Int64Index`` and ``Float64Index`` preventing the use of the modulo operator (:issue:`9244`)
 
+
 - Bug in ``DataFrame`` when masking an empty ``DataFrame`` (:issue:`11859`)
 
-- Removed ``millisecond`` property of ``DatetimeIndex``. This would always raise
-  a ``ValueError`` (:issue:`12019`).
+
+
+
+
+
+
+
+- Removed ``millisecond`` property of ``DatetimeIndex``. This would always raise a ``ValueError`` (:issue:`12019`).
+- Bug in ``Series`` constructor with read-only data (:issue:`11502`)

pandas/src/generate_code.py

@@ -88,13 +88,7 @@
 """
 
 
-take_1d_template = """
-@cython.wraparound(False)
-@cython.boundscheck(False)
-def take_1d_%(name)s_%(dest)s(%(c_type_in)s[:] values,
-                              int64_t[:] indexer,
-                              %(c_type_out)s[:] out,
-                              fill_value=np.nan):
+inner_take_1d_template = """\
     cdef:
         Py_ssize_t i, n, idx
         %(c_type_out)s fv
@@ -112,6 +106,33 @@ def take_1d_%(name)s_%(dest)s(%(c_type_in)s[:] values,
     %(tab)s        out[i] = %(preval)svalues[idx]%(postval)s
 """
 
+take_1d_template = """\
+@cython.wraparound(False)
+@cython.boundscheck(False)
+cdef inline take_1d_%(name)s_%(dest)s_memview(%(c_type_in)s[:] values,
+                      int64_t[:] indexer,
+                      %(c_type_out)s[:] out,
+                      fill_value=np.nan):
+""" + inner_take_1d_template + """
+
+@cython.wraparound(False)
+@cython.boundscheck(False)
+def take_1d_%(name)s_%(dest)s(ndarray[%(c_type_in)s, ndim=1] values,
+                              int64_t[:] indexer,
+                              %(c_type_out)s[:] out,
+                              fill_value=np.nan):
+
+    if values.flags.writeable:
+        # We can call the memoryview version of the code
+        take_1d_%(name)s_%(dest)s_memview(values, indexer, out,
+                                          fill_value=fill_value)
+        return
+
+    # We cannot use the memoryview version on readonly-buffers due to
+    # a limitation of Cython's typed memoryviews. Instead we can use
+    # the slightly slower Cython ndarray type directly.
+""" + inner_take_1d_template
+
 inner_take_2d_axis0_template = """\
     cdef:
         Py_ssize_t i, j, k, n, idx