Update pandas version to 0.24 (#451)

* Update pandas version to 0.24

* pandas release notes: http://pandas.pydata.org/pandas-docs/stable/whatsnew/v0.24.0.html
* Update imports to match changes in pandas
* Add functionality for list of functions on `axis=1` for `apply`
* Remove `pd.match` from API
* Small regression in pandas requires regression in Modin
  * pandas-dev/pandas#25101 reports this issue
  * pandas-dev/pandas#25102 resolves this issue
* TODO: Expose `pandas.Array` once we properly test

* Finishing regression update in `all`/`any`

* Update to pandas 0.24 in setup.py and requirements.txt

* Bump to 0.24.1

* Update API and add a test for the API

* Add test for API, update API

* Update API test and finalize compatibility updates

* Revert bug

* Cleanup and add tests

* Fix bug in test

* Lint

* Lint

* Remove print

* Fix transform tests and bug in transform

* Add list test for test_rename

* Fix transform bug

.coveragerc

@@ -20,4 +20,6 @@ exclude_lines =
     pragma: no cover
     # Don't complain if tests don't hit defensive assertion code:
     raise AssertionError
-    raise NotImplementedError
+    raise NotImplementedError
+    raise ImportError
+    assert

.travis.yml

@@ -40,6 +40,15 @@ matrix:
         - black --check modin/
         - flake8 .
 
+    - os: linux
+      dist: trusty
+      env:
+        - PYTHON=3.6
+        - API_COMPAT=1
+      script:
+        - export PATH="$HOME/miniconda/bin:$PATH"
+        - python -m pytest modin/pandas/test/test_api.py
+
 install:
   - ./.travis/install-dependencies.sh
 

modin/data_management/query_compiler/pandas_query_compiler.py

@@ -13,7 +13,7 @@
     is_datetime_or_timedelta_dtype,
     is_bool_dtype,
 )
-from pandas.core.index import _ensure_index
+from pandas.core.index import ensure_index
 from pandas.core.base import DataError
 
 from modin.engines.base.block_partitions import BaseBlockPartitions
@@ -97,7 +97,7 @@ def pandas_index_extraction(df, axis):
         return index_obj[new_indices] if compute_diff else new_indices
 
     def _validate_set_axis(self, new_labels, old_labels):
-        new_labels = _ensure_index(new_labels)
+        new_labels = ensure_index(new_labels)
         old_len = len(old_labels)
         new_len = len(new_labels)
         if old_len != new_len:
@@ -118,14 +118,14 @@ def _get_columns(self):
 
     def _set_index(self, new_index):
         if self._index_cache is None:
-            self._index_cache = _ensure_index(new_index)
+            self._index_cache = ensure_index(new_index)
         else:
             new_index = self._validate_set_axis(new_index, self._index_cache)
             self._index_cache = new_index
 
     def _set_columns(self, new_columns):
         if self._columns_cache is None:
-            self._columns_cache = _ensure_index(new_columns)
+            self._columns_cache = ensure_index(new_columns)
         else:
             new_columns = self._validate_set_axis(new_columns, self._columns_cache)
             self._columns_cache = new_columns
@@ -1388,11 +1388,16 @@ def _process_all_any(self, func, **kwargs):
 
         if bool_only:
             if axis == 0 and not axis_none and len(not_bool_col) == len(self.columns):
-                return pandas.Series(dtype=bool)
-            if len(not_bool_col) == len(self.columns):
-                query_compiler = self
-            else:
-                query_compiler = self.drop(columns=not_bool_col)
+                # TODO add this line back once pandas-dev/pandas#25101 is resolved
+                # return pandas.Series(dtype=bool)
+                pass
+            # See note above about pandas-dev/pandas#25101
+            # TODO remove this when pandas 0.24.2 is released.
+            query_compiler = self
+            # if len(not_bool_col) == len(self.columns):
+            #     query_compiler = self
+            # else:
+            #     query_compiler = self.drop(columns=not_bool_col)
         else:
             if (
                 bool_only is False
@@ -2492,11 +2497,22 @@ def _list_like_func(self, func, axis, *args, **kwargs):
         Returns:
             A new PandasQueryCompiler.
         """
-        func_prepared = self._prepare_method(lambda df: df.apply(func, *args, **kwargs))
+        func_prepared = self._prepare_method(
+            lambda df: df.apply(func, axis, *args, **kwargs)
+        )
         new_data = self._map_across_full_axis(axis, func_prepared)
-        # When the function is list-like, the function names become the index
-        new_index = [f if isinstance(f, string_types) else f.__name__ for f in func]
-        return self.__constructor__(new_data, new_index, self.columns)
+        # When the function is list-like, the function names become the index/columns
+        new_index = (
+            [f if isinstance(f, string_types) else f.__name__ for f in func]
+            if axis == 0
+            else self.index
+        )
+        new_columns = (
+            [f if isinstance(f, string_types) else f.__name__ for f in func]
+            if axis == 1
+            else self.columns
+        )
+        return self.__constructor__(new_data, new_index, new_columns)
 
     def _callable_func(self, func, axis, *args, **kwargs):
         """Apply callable functions across given axis.

modin/engines/base/io.py

@@ -435,6 +435,7 @@ def to_sql(
         index_label=None,
         chunksize=None,
         dtype=None,
+        method=None,
     ):
         ErrorMessage.default_to_pandas("`to_sql`")
         df = qc.to_pandas()
@@ -447,4 +448,5 @@ def to_sql(
             index_label=index_label,
             chunksize=chunksize,
             dtype=dtype,
+            method=method,
         )

modin/pandas/__init__.py

@@ -14,7 +14,6 @@
     factorize,
     test,
     qcut,
-    match,
     Panel,
     date_range,
     period_range,
@@ -64,7 +63,7 @@
 from .plotting import Plotting as plotting
 from .. import __execution_engine__ as execution_engine
 
-__pandas_version__ = "0.23.4"
+__pandas_version__ = "0.24.1"
 
 if pandas.__version__ != __pandas_version__:
     raise ImportError(
@@ -131,7 +130,7 @@ def initialize_ray():
 if execution_engine == "Ray":
     initialize_ray()
     num_cpus = ray.global_state.cluster_resources()["CPU"]
-elif execution_engine == "Dask":
+elif execution_engine == "Dask":  # pragma: no cover
     from distributed.client import _get_global_client
 
     if threading.current_thread().name == "MainThread":
@@ -174,7 +173,6 @@ def initialize_ray():
     "factorize",
     "test",
     "qcut",
-    "match",
     "to_datetime",
     "get_dummies",
     "isna",