From 8bac66c94fc2ea67becd8f54d51cfd8bd00e4c48 Mon Sep 17 00:00:00 2001
From: Martin Fleischmann <martin@martinfleischmann.net>
Date: Sun, 5 Jun 2022 09:26:07 +0200
Subject: [PATCH] REF: deprecate GeoDataFrame._crs (#2373)

Co-authored-by: Joris Van den Bossche <jorisvandenbossche@gmail.com>
---
 geopandas/geodataframe.py                     |  78 +++++++++++++-----
 ...5.1_pd-0.25.3_py-3.7.3_x86_64_linux.pickle | Bin 1542 -> 0 bytes
 geopandas/tests/test_crs.py                   |  37 ++++++++-
 geopandas/tests/test_merge.py                 |   1 +
 4 files changed, 92 insertions(+), 24 deletions(-)
 delete mode 100644 geopandas/io/tests/data/pickle/0.5.1_pd-0.25.3_py-3.7.3_x86_64_linux.pickle

diff --git a/geopandas/geodataframe.py b/geopandas/geodataframe.py
index 48bb03e..a13f1ed 100644
--- a/geopandas/geodataframe.py
+++ b/geopandas/geodataframe.py
@@ -49,7 +49,9 @@ def _ensure_geometry(data, crs=None):
     if is_geometry_type(data):
         if isinstance(data, Series):
             data = GeoSeries(data)
-        if data.crs is None:
+        if data.crs is None and crs is not None:
+            # Avoids caching issues/crs sharing issues
+            data = data.copy()
             data.crs = crs
         return data
     else:
@@ -121,6 +123,7 @@ class GeoDataFrame(GeoPandasBase, DataFrame):
     GeoSeries : Series object designed to store shapely geometry objects
     """
 
+    # TODO: remove "_crs" in 0.12
     _metadata = ["_crs", "_geometry_column_name"]
 
     _geometry_column_name = DEFAULT_GEO_COLUMN_NAME
@@ -129,9 +132,8 @@ class GeoDataFrame(GeoPandasBase, DataFrame):
         with compat.ignore_shapely2_warnings():
             super().__init__(data, *args, **kwargs)
 
-        # need to set this before calling self['geometry'], because
-        # getitem accesses crs
-        self._crs = CRS.from_user_input(crs) if crs else None
+        # TODO: to be removed in 0.12
+        self._crs = None
 
         # set_geometry ensures the geometry data have the proper dtype,
         # but is not called if `geometry=None` ('geometry' column present
@@ -180,7 +182,7 @@ class GeoDataFrame(GeoPandasBase, DataFrame):
             ):
                 raise ValueError(crs_mismatch_error)
 
-            self.set_geometry(geometry, inplace=True)
+            self.set_geometry(geometry, inplace=True, crs=crs)
 
         if geometry is None and crs:
             raise ValueError(
@@ -330,8 +332,7 @@ class GeoDataFrame(GeoPandasBase, DataFrame):
             del frame[to_remove]
 
         if not crs:
-            level_crs = getattr(level, "crs", None)
-            crs = level_crs if level_crs is not None else self._crs
+            crs = getattr(level, "crs", None)
 
         if isinstance(level, (GeoSeries, GeometryArray)) and level.crs != crs:
             # Avoids caching issues/crs sharing issues
@@ -342,7 +343,9 @@ class GeoDataFrame(GeoPandasBase, DataFrame):
         level = _ensure_geometry(level, crs=crs)
         frame[geo_column_name] = level
         frame._geometry_column_name = geo_column_name
-        frame.crs = crs
+
+        # TODO: to be removed in 0.12
+        frame._crs = level.crs
         if not inplace:
             return frame
 
@@ -424,7 +427,19 @@ class GeoDataFrame(GeoPandasBase, DataFrame):
         GeoDataFrame.to_crs : re-project to another CRS
 
         """
-        return self._crs
+        # TODO: remove try/except in 0.12
+        try:
+            return self.geometry.crs
+        except AttributeError:
+            # the active geometry column might not be set
+            warnings.warn(
+                "Accessing CRS of a GeoDataFrame without a geometry column is "
+                "deprecated and will be removed in GeoPandas 0.12. "
+                "Use GeoDataFrame.set_geometry to set the active geometry column.",
+                FutureWarning,
+                stacklevel=2,
+            )
+            return self._crs
 
     @crs.setter
     def crs(self, value):
@@ -438,20 +453,27 @@ class GeoDataFrame(GeoPandasBase, DataFrame):
         else:
             if hasattr(self.geometry.values, "crs"):
                 self.geometry.values.crs = value
-                self._crs = self.geometry.values.crs
             else:
                 # column called 'geometry' without geometry
                 self._crs = None if not value else CRS.from_user_input(value)
 
+                # TODO: raise this error in 0.12. This already raises a FutureWarning
+                # TODO: defined in the crs property above
+                # raise ValueError(
+                #     "Assigning CRS to a GeoDataFrame without an active geometry "
+                #     "column is not supported. Use GeoDataFrame.set_geometry to set "
+                #     "the active geometry column.",
+                # )
+
     def __setstate__(self, state):
         # overriding DataFrame method for compat with older pickles (CRS handling)
+        crs = None
         if isinstance(state, dict):
-            if "_metadata" in state and "crs" in state["_metadata"]:
-                metadata = state["_metadata"]
-                metadata[metadata.index("crs")] = "_crs"
             if "crs" in state and "_crs" not in state:
-                crs = state.pop("crs")
-                state["_crs"] = CRS.from_user_input(crs) if crs is not None else crs
+                crs = state.pop("crs", None)
+            else:
+                crs = state.pop("_crs", None)
+            crs = CRS.from_user_input(crs) if crs is not None else crs
 
         super().__setstate__(state)
 
@@ -459,9 +481,9 @@ class GeoDataFrame(GeoPandasBase, DataFrame):
         # at GeoDataFrame level with '_crs' (and not 'crs'), so without propagating
         # to the GeoSeries/GeometryArray
         try:
-            if self.crs is not None:
+            if crs is not None:
                 if self.geometry.values.crs is None:
-                    self.crs = self.crs
+                    self.crs = crs
         except Exception:
             pass
 
@@ -1316,7 +1338,6 @@ individually so that features may have different properties
             df = self.copy()
         geom = df.geometry.to_crs(crs=crs, epsg=epsg)
         df.geometry = geom
-        df.crs = geom.crs
         if not inplace:
             return df
 
@@ -1377,7 +1398,6 @@ individually so that features may have different properties
                     result._geometry_column_name = geo_col
                 else:
                     result._geometry_column_name = None
-                    result._crs = None
             else:
                 result.__class__ = DataFrame
         return result
@@ -1391,8 +1411,24 @@ individually so that features may have different properties
             if pd.api.types.is_scalar(value) or isinstance(value, BaseGeometry):
                 value = [value] * self.shape[0]
             try:
-                value = _ensure_geometry(value, crs=self.crs)
-                self._crs = value.crs
+                # TODO: remove this use of _crs in 0.12
+                warn = False
+                if not (hasattr(self, "geometry") and hasattr(self.geometry, "crs")):
+                    crs = self._crs
+                    warn = True
+                else:
+                    crs = getattr(self, "crs", None)
+                value = _ensure_geometry(value, crs=crs)
+                if warn and crs is not None:
+                    warnings.warn(
+                        "Setting geometries to a GeoDataFrame without a geometry "
+                        "column will currently preserve the CRS, if present. "
+                        "This is deprecated, and in the future the CRS will be lost "
+                        "in this case. You can use set_crs(..) on the result to "
+                        "set the CRS manually.",
+                        FutureWarning,
+                        stacklevel=2,
+                    )
             except TypeError:
                 warnings.warn("Geometry column does not contain geometry.")
         super().__setitem__(key, value)
diff --git a/geopandas/io/tests/data/pickle/0.5.1_pd-0.25.3_py-3.7.3_x86_64_linux.pickle b/geopandas/io/tests/data/pickle/0.5.1_pd-0.25.3_py-3.7.3_x86_64_linux.pickle
deleted file mode 100644
index 1d97ad597a4b064da16b16741709e4e365a906a6..0000000000000000000000000000000000000000
GIT binary patch
literal 0
HcmV?d00001

literal 1542
zcmdUuOHUL*5P)ZPXJ8>1Vhji|F(w{ZIgIj%QDa;pD8Y#%!Lyn4%r;va=CPRvuqX*0
zfSB~f&>qxZ;!p7GAK~snt9oV)tV{4}AEv9Py1HIp9T=Z}nwnbsvM$D2%}qB0?zVUw
zN->FBheb`6S|~MH#x@fcaw*1EdAx$h=jwPO@4S~}Zn|_wikqd8A1B;uM;T8d7Nl0l
zBG%$biuO-~*x!0toTaj_zrhEl&KoxQE>A6wrRZ|hjrfkc7U4fMfGfi6$>iB698W}f
z*y&ozZ<q&Jn<WYBqCYp`ejKG)lKUB&3sYT`Vi=QCS5Bm@9m-|hlgDyH)@|Lc97?z@
z%x2c@a5Aa9dVqIqz0g;M{!TiM^WqmS^|2e7^gd7)I+mdJr-~eZ3nQYsIr$Mj4&!2a
zP@V~kT9oJp?PrpMSx1;@#*z&E3nR_qj<l;Ja1?a)CP%+83DcAW+~}pC+tnipw?w7g
z7y9GA&@~Euc}g!x&?@C#Eq96Pi(bF2Ql3nKb$HMvPC}j~U8@sg<FU}|YR49D$mI#W
z=Rwm>8^z<N4Saj8U#l&Zu9pUBYo)e^S}L!~ejKhM32<#dpjZ<M6J{7CGTUpz<p%Y$
z27F(Np?lWcqBW;zHyF|>VN48pL~|;!2N;1E+Gascnunqmwh|YWh<we<+X!TYqL~J5
zpO_+?3ZQZdeG66~fSA}T|MEP=h=t)Igi?$U<mCj8xL9VUs7JS=>eE<FsJW%`yM7$x
zVdN_8IP5v_7V5bthH1d*E*Qnk!Z3`YMAHWsSCD2wH5z6_t=D2nr9V^6yOhJM19P?p
z^L7;${*BGeb8Nzs^FvjV`Q<aMZ;fLA-dIrUN8?g5|3>S-0PnaDJV5yB(1HU1>=emc
n@KM3ikI>hF2*Eb&z%F!QPmDZUf4%x>abf<xL@17PFSmXJUc4CH

diff --git a/geopandas/tests/test_crs.py b/geopandas/tests/test_crs.py
index 93eb6b4..a587e5a 100644
--- a/geopandas/tests/test_crs.py
+++ b/geopandas/tests/test_crs.py
@@ -209,7 +209,6 @@ class TestGeometryArrayCRS:
             s = GeoSeries(arr, crs=4326)
         assert s.crs == self.osgb
 
-    @pytest.mark.filterwarnings("ignore:Assigning CRS")
     def test_dataframe(self):
         arr = from_shapely(self.geoms, crs=27700)
         df = GeoDataFrame(geometry=arr)
@@ -298,8 +297,23 @@ class TestGeometryArrayCRS:
 
         # geometry column without geometry
         df = GeoDataFrame({"geometry": [0, 1]})
-        df.crs = 27700
-        assert df.crs == self.osgb
+        with pytest.warns(
+            FutureWarning, match="Accessing CRS of a GeoDataFrame without a geometry"
+        ):
+            df.crs = 27700
+        with pytest.warns(
+            FutureWarning, match="Accessing CRS of a GeoDataFrame without a geometry"
+        ):
+            assert df.crs == self.osgb
+
+    def test_dataframe_getitem_without_geometry_column(self):
+        df = GeoDataFrame({"col": range(10)}, geometry=self.arr)
+        df["geom2"] = df.geometry.centroid
+        subset = df[["col", "geom2"]]
+        with pytest.warns(
+            FutureWarning, match="Accessing CRS of a GeoDataFrame without a geometry"
+        ):
+            assert subset.crs == self.osgb
 
     def test_dataframe_setitem(self):
         # new geometry CRS has priority over GDF CRS
@@ -337,6 +351,22 @@ class TestGeometryArrayCRS:
         assert df["geometry"].crs == self.wgs
         assert df["other_geom"].crs == self.osgb
 
+    def test_dataframe_setitem_without_geometry_column(self):
+        arr = from_shapely(self.geoms)
+        df = GeoDataFrame({"col1": [1, 2], "geometry": arr}, crs=4326)
+
+        # create a dataframe without geometry column, but currently has cached _crs
+        with pytest.warns(UserWarning):
+            df["geometry"] = 1
+
+        # assigning a list of geometry object will currently use _crs
+        with pytest.warns(
+            FutureWarning,
+            match="Setting geometries to a GeoDataFrame without a geometry",
+        ):
+            df["geometry"] = self.geoms
+        assert df.crs == self.wgs
+
     @pytest.mark.parametrize(
         "scalar", [None, Point(0, 0), LineString([(0, 0), (1, 1)])]
     )
@@ -348,6 +378,7 @@ class TestGeometryArrayCRS:
         assert df.geometry.crs == self.wgs
         assert df.geometry.values.crs == self.wgs
 
+    @pytest.mark.filterwarnings("ignore:Accessing CRS")
     def test_crs_with_no_geom_fails(self):
         with pytest.raises(ValueError, match="Assigning CRS to a GeoDataFrame without"):
             df = GeoDataFrame()
diff --git a/geopandas/tests/test_merge.py b/geopandas/tests/test_merge.py
index a29fbae..0de01f3 100644
--- a/geopandas/tests/test_merge.py
+++ b/geopandas/tests/test_merge.py
@@ -105,6 +105,7 @@ class TestMerging:
         reason="pandas <=1.0 hard codes concat([GeoSeries, GeoSeries]) -> "
         "DataFrame or Union[DataFrame, SparseDataFrame] in 0.25",
     )
+    @pytest.mark.filterwarnings("ignore:Accessing CRS")
     def test_concat_axis1_geoseries(self):
         gseries2 = GeoSeries([Point(i, i) for i in range(3, 6)], crs="epsg:4326")
         result = pd.concat([gseries2, self.gseries], axis=1)