Raise ParserError instead of IndexError when specifying an incorrect number of columns with index_col for the read_csv C parser.

Author: DriesSchaumont

doc/source/whatsnew/v1.6.0.rst

@@ -35,7 +35,7 @@ Other enhancements
 - Added ``index`` parameter to :meth:`DataFrame.to_dict` (:issue:`46398`)
 - Added metadata propagation for binary operators on :class:`DataFrame` (:issue:`28283`)
 - :class:`.CategoricalConversionWarning`, :class:`.InvalidComparison`, :class:`.InvalidVersion`, :class:`.LossySetitemError`, and :class:`.NoBufferPresent` are now exposed in ``pandas.errors`` (:issue:`27656`)
--
+- :func:`read_csv`: specifying an incorrect number of columns with ``index_col`` of now raises ``ParserError`` instead of ``IndexError`` when using the c parser.
 
 .. ---------------------------------------------------------------------------
 .. _whatsnew_160.notable_bug_fixes:

pandas/io/parsers/c_parser_wrapper.py

@@ -33,6 +33,7 @@
 
 from pandas.io.parsers.base_parser import (
     ParserBase,
+    ParserError,
     is_index_col,
 )
 
@@ -270,6 +271,13 @@ def read(
             # implicit index, no index names
             arrays = []
 
+            if self.index_col and self._reader.leading_cols != len(self.index_col):
+                raise ParserError(
+                    "Could not construct index. Requested to use "
+                    f"{len(self.index_col)} number of columns, but "
+                    f"{self._reader.leading_cols} left to parse."
+                )
+
             for i in range(self._reader.leading_cols):
                 if self.index_col is None:
                     values = data.pop(i)

pandas/tests/io/parser/common/test_read_errors.py

@@ -292,6 +292,18 @@ def test_conflict_on_bad_line(all_parsers, error_bad_lines, warn_bad_lines):
         parser.read_csv(StringIO(data), on_bad_lines="error", **kwds)
 
 
+def test_bad_header_uniform_error(all_parsers):
+    parser = all_parsers
+    data = "+++123456789...\ncol1,col2,col3,col4\n1,2,3,4\n"
+    msg = "Expected 2 fields in line 2, saw 4"
+    if parser.engine == "c":
+        msg = "Could not construct index. Requested to use 1 "
+        "number of columns, but 3 left to parse."
+
+    with pytest.raises(ParserError, match=msg):
+        parser.read_csv(StringIO(data), index_col=0, on_bad_lines="error")
+
+
 def test_on_bad_lines_warn_correct_formatting(all_parsers, capsys):
     # see gh-15925
     parser = all_parsers