fix(cli): --zero breaks open-ended column ranges

CHANGELOG.rst

@@ -5,6 +5,7 @@ Unreleased
 -  feat: :doc:`/scripts/csvclean` adds a :code:`--remove-empty-columns` option to remove empty columns from standard output.
 -  feat: :doc:`/scripts/in2csv` guesses the ``ndjson`` format for files with :code:`.ndjson`, :code:`.jsonl` and :code:`.jl` extensions.
 -  fix: :code:`-C/--not-columns` now excludes the last column of an open-ended range (e.g. :code:`2-`).
+-  fix: An open-ended range (e.g. :code:`0:` or :code:`:1`) now respects the :code:`--zero` option.
 
 2.2.0 - December 15, 2025
 -------------------------

csvkit/cli.py

@@ -534,8 +534,10 @@ def _resolve_column_identifier(identifier, column_names, column_offset, ignore_u
             raise
 
         try:
-            a = int(a) if a else 1
-            b = int(b) + 1 if b else len(column_names) + 1
+            # The bounds are in identifier space, so the implicit ends of an open-ended range
+            # depend on whether identifiers are 1-based (the default) or 0-based (--zero).
+            a = int(a) if a else column_offset
+            b = int(b) + 1 if b else len(column_names) + column_offset
         except ValueError:
             if ignore_invalid_range:
                 return []

tests/test_cli.py

@@ -84,9 +84,20 @@ def test_exclude_ignores_unknown_names_but_reports_invalid_ranges(self):
 
     def test_range_notation_open_ended(self):
         self.assertEqual([0, 1, 2], parse_column_identifiers(':3', self.headers))
+        self.assertEqual([0, 1, 2, 3], parse_column_identifiers(':3', self.headers, column_offset=0))
 
         target = list(range(3, len(self.headers)))  # protect against devs adding to self.headers
         target.insert(0, 0)
         self.assertEqual(target, parse_column_identifiers('1,4:', self.headers))
 
+        target = list(range(4, len(self.headers)))  # protect against devs adding to self.headers
+        target.insert(0, 1)
+        self.assertEqual(target, parse_column_identifiers('1,4:', self.headers, column_offset=0))
+
         self.assertEqual(list(range(0, len(self.headers))), parse_column_identifiers('1:', self.headers))
+        self.assertEqual(list(range(0, len(self.headers))),
+                         parse_column_identifiers('0:', self.headers, column_offset=0))
+
+        # An open-ended exclusion range respects the column offset too.
+        self.assertEqual([0], parse_column_identifiers(None, self.headers, excluded_columns='2-'))
+        self.assertEqual([0], parse_column_identifiers(None, self.headers, column_offset=0, excluded_columns='1-'))