ERR: improves message raised for bad names in usecols (GH14154)

doc/source/whatsnew/v0.19.0.txt

@@ -471,6 +471,7 @@ Other enhancements
 - :meth:`~DataFrame.to_html` now has a ``border`` argument to control the value in the opening ``<table>`` tag. The default is the value of the ``html.border`` option, which defaults to 1. This also affects the notebook HTML repr, but since Jupyter's CSS includes a border-width attribute, the visual effect is the same. (:issue:`11563`).
 - Raise ``ImportError`` in the sql functions when ``sqlalchemy`` is not installed and a connection string is used (:issue:`11920`).
 - Compatibility with matplotlib 2.0. Older versions of pandas should also work with matplotlib 2.0 (:issue:`13333`)
+- When using the ``usecols`` argument in the ``read`` functions, specifying a column name that isn't found now generates a more helpful error message (:issue:`14154`)
 
 .. _whatsnew_0190.api:
 

pandas/io/parsers.py

@@ -981,8 +981,7 @@ def _validate_usecols_arg(usecols):
 
     if usecols is not None:
         usecols_dtype = lib.infer_dtype(usecols)
-        if usecols_dtype not in ('empty', 'integer',
-                                 'string', 'unicode'):
+        if usecols_dtype not in ('empty', 'integer', 'string', 'unicode'):
             raise ValueError(msg)
 
         return set(usecols)
@@ -1424,7 +1423,13 @@ def __init__(self, src, **kwds):
                               if (i in self.usecols or n in self.usecols)]
 
             if len(self.names) < len(self.usecols):
-                raise ValueError("Usecols do not match names.")
+                bad_cols = [n for n in self.usecols if n not in self.names]
+                if len(bad_cols) > 0:
+                    raise ValueError(("%s specified in usecols but not found "
+                                     "in names.") % bad_cols)
+                else:
+                    raise ValueError(("Number of usecols is greater than "
+                                      "number of names."))
 
         self._set_noconvert_columns()
 
@@ -2185,16 +2190,21 @@ def _handle_usecols(self, columns, usecols_key):
         usecols_key is used if there are string usecols.
         """
         if self.usecols is not None:
-            if any([isinstance(u, string_types) for u in self.usecols]):
+            if any([isinstance(c, string_types) for c in self.usecols]):
                 if len(columns) > 1:
                     raise ValueError("If using multiple headers, usecols must "
                                      "be integers.")
+                bad_cols = [n for n in self.usecols if n not in usecols_key]
+                if len(bad_cols) > 0:
+                    raise ValueError(("%s specified in usecols but not found "
+                                     "in names.") % bad_cols)
+
                 col_indices = []
-                for u in self.usecols:
-                    if isinstance(u, string_types):
-                        col_indices.append(usecols_key.index(u))
+                for c in self.usecols:
+                    if isinstance(c, string_types):
+                        col_indices.append(usecols_key.index(c))
                     else:
-                        col_indices.append(u)
+                        col_indices.append(c)
             else:
                 col_indices = self.usecols
 

pandas/io/tests/parser/usecols.py

@@ -83,6 +83,8 @@ def test_usecols(self):
         # length conflict, passed names and usecols disagree
         self.assertRaises(ValueError, self.read_csv, StringIO(data),
                           names=['a', 'b'], usecols=[1], header=None)
+        self.assertRaises(ValueError, self.read_csv, StringIO(data),
+                          names=['a', 'b'], usecols=['A'], header=None)
 
     def test_usecols_index_col_False(self):
         # see gh-9082