Deprecated usecols with out of bounds indices in read_csv

doc/source/whatsnew/v1.3.0.rst

@@ -797,6 +797,7 @@ I/O
 - Bug in :func:`read_excel` raising ``AttributeError`` with ``MultiIndex`` header followed by two empty rows and no index, and bug affecting :func:`read_excel`, :func:`read_csv`, :func:`read_table`, :func:`read_fwf`, and :func:`read_clipboard` where one blank row after a ``MultiIndex`` header with no index would be dropped (:issue:`40442`)
 - Bug in :meth:`DataFrame.to_string` misplacing the truncation column when ``index=False`` (:issue:`40907`)
 - Bug in :func:`read_orc` always raising ``AttributeError`` (:issue:`40918`)
+- Bug in :func:`read_csv` raising uncontrolled ``ValueError`` when ``usecols`` index is ouf of bounds, now raising ``ParserError`` (:issue:`25623`)
 
 Period
 ^^^^^^

pandas/io/parsers/python_parser.py

@@ -470,12 +470,14 @@ def _infer_columns(self):
                 if self.usecols is not None:
                     # Set _use_cols. We don't store columns because they are
                     # overwritten.
-                    self._handle_usecols(columns, names)
+                    self._handle_usecols(columns, names, num_original_columns)
                 else:
                     num_original_columns = len(names)
                 columns = [names]
             else:
-                columns = self._handle_usecols(columns, columns[0])
+                columns = self._handle_usecols(
+                    columns, columns[0], num_original_columns
+                )
         else:
             try:
                 line = self._buffered_line()
@@ -494,10 +496,12 @@ def _infer_columns(self):
                     columns = [[f"{self.prefix}{i}" for i in range(ncols)]]
                 else:
                     columns = [list(range(ncols))]
-                columns = self._handle_usecols(columns, columns[0])
+                columns = self._handle_usecols(
+                    columns, columns[0], num_original_columns
+                )
             else:
                 if self.usecols is None or len(names) >= num_original_columns:
-                    columns = self._handle_usecols([names], names)
+                    columns = self._handle_usecols([names], names, num_original_columns)
                     num_original_columns = len(names)
                 else:
                     if not callable(self.usecols) and len(names) != len(self.usecols):
@@ -506,13 +510,13 @@ def _infer_columns(self):
                             "header fields in the file"
                         )
                     # Ignore output but set used columns.
-                    self._handle_usecols([names], names)
+                    self._handle_usecols([names], names, ncols)
                     columns = [names]
                     num_original_columns = ncols
 
         return columns, num_original_columns, unnamed_cols
 
-    def _handle_usecols(self, columns, usecols_key):
+    def _handle_usecols(self, columns, usecols_key, num_original_columns):
         """
         Sets self._col_indices
 
@@ -537,6 +541,13 @@ def _handle_usecols(self, columns, usecols_key):
                     else:
                         col_indices.append(col)
             else:
+                missing_usecols = [
+                    col for col in self.usecols if col >= num_original_columns
+                ]
+                if missing_usecols:
+                    raise ParserError(
+                        f"Usecols indices {missing_usecols} are out of bounds!"
+                    )
                 col_indices = self.usecols
 
             columns = [

pandas/tests/io/parser/test_python_parser_only.py

@@ -312,3 +312,19 @@ def test_malformed_skipfooter(python_parser_only):
     msg = "Expected 3 fields in line 4, saw 5"
     with pytest.raises(ParserError, match=msg):
         parser.read_csv(StringIO(data), header=1, comment="#", skipfooter=1)
+
+
+@pytest.mark.parametrize("header", [0, None])
+@pytest.mark.parametrize("names", [None, ["a", "b"], ["a", "b", "c"]])
+def test_usecols_indices_out_of_bounds(python_parser_only, names, header):
+    # GH#25623
+    if header == 0 and names == ["a", "b", "c"]:
+        pytest.skip("This case is not valid")
+    parser = python_parser_only
+    data = """
+    a,b
+    1,2
+    """
+    msg = r"Usecols indices \[2\] are out of bounds!"
+    with pytest.raises(ParserError, match=msg):
+        parser.read_csv(StringIO(data), usecols=[0, 2], names=names, header=header)