Fix issue when REAL is already present in loaded CSV files (#67)

src/fmu/ensemble/realization.py

@@ -384,6 +384,15 @@ def load_csv(self, localpath, convert_numeric=True, force_reread=False):
                 else:
                     dtype = str
                 dframe = pd.read_csv(fullpath, dtype=dtype)
+                if "REAL" in dframe:
+                    dframe.rename(columns={"REAL": "REAL_ORIG"}, inplace=True)
+                    logger.warning(
+                        (
+                            "Loaded file %s already had the column REAL, "
+                            "this was renamed to REAL_ORIG"
+                        ),
+                        fullpath,
+                    )
             except pd.errors.EmptyDataError:
                 dframe = None  # or empty dataframe?
 

tests/test_realization.py

@@ -139,6 +139,13 @@ def test_single_realization():
     # test basename and no extension:
     assert isinstance(real.get_df("simulator_volume_fipnum"), pd.DataFrame)
 
+    # Some CSV files might already contain the REAL column, this is not allowed
+    foo_df = pd.DataFrame(columns=["REAL", "FOOBAR"], data=[[0, 1], [2, 3]])
+    foo_df.to_csv(os.path.join(realdir, "foo-real.csv"), index=False)
+    real.load_csv("foo-real.csv")  # A warning will be issued.
+    assert "REAL" not in real.get_df("foo-real")
+    assert "FOOBAR" in real.get_df("foo-real")
+
     with pytest.raises(ValueError):
         real.get_df("notexisting.csv")