BUG: Fix read_csv IndexError crash for c engine with header=None and 2 (or more) extra columns (pandas-dev#32839)

Author: roberthdevries

doc/source/whatsnew/v1.1.0.rst

@@ -349,6 +349,7 @@ I/O
 - Bug in :meth:`read_csv` was causing a file descriptor leak on an empty file (:issue:`31488`)
 - Bug in :meth:`read_csv` was causing a segfault when there were blank lines between the header and data rows (:issue:`28071`)
 - Bug in :meth:`read_csv` was raising a misleading exception on a permissions issue (:issue:`23784`)
+- Bug in :meth:`read_csv` was raising an ``IndexError`` when header=None and 2 extra data columns
 
 
 Plotting

pandas/_libs/parsers.pyx

@@ -1316,8 +1316,8 @@ cdef class TextReader:
         else:
             if self.header is not None:
                 j = i - self.leading_cols
-                # hack for #2442
-                if j == len(self.header[0]):
+                # generate extra (bogus) headers if there are more columns than headers
+                if j >= len(self.header[0]):
                     return j
                 else:
                     return self.header[0][j]

pandas/tests/io/parser/test_common.py

@@ -2116,3 +2116,13 @@ def test_blank_lines_between_header_and_data_rows(all_parsers, nrows):
     parser = all_parsers
     df = parser.read_csv(StringIO(csv), header=3, nrows=nrows, skip_blank_lines=False)
     tm.assert_frame_equal(df, ref[:nrows])
+
+
+def test_no_header_two_extra_columns(all_parsers):
+    # GH 26218
+    column_names = ["one", "two", "three"]
+    ref = DataFrame([["foo", "bar", "baz"]], columns=column_names)
+    stream = StringIO("foo,bar,baz,bam,blah")
+    parser = all_parsers
+    df = parser.read_csv(stream, header=None, names=column_names, index_col=False)
+    tm.assert_frame_equal(df, ref)