Improve parsing of PEP249-style placeholder (fixes andialbrecht#103).

andialbrecht · andialbrecht · commit 84725ee64570 · 2013-06-29T19:14:58.000+02:00
diff --git a/CHANGES b/CHANGES
@@ -13,6 +13,7 @@ Enhancements
 * sqlparse.split() now removes leading and trailing whitespaces from splitted
   statements.
 * Add USE as keyword token (by mulos).
+* Improve parsing of PEP249-style placeholders (issue103).
 
 
 Release 0.1.7 (Apr 06, 2013)
diff --git a/sqlparse/lexer.py b/sqlparse/lexer.py
@@ -182,6 +182,7 @@ class Lexer(object):
             (r"´(´´|[^´])*´", tokens.Name),
             (r'\$([^\W\d]\w*)?\$', tokens.Name.Builtin),
             (r'\?{1}', tokens.Name.Placeholder),
+            (r'%\(\w+\)s', tokens.Name.Placeholder),
             (r'[$:?%]\w+', tokens.Name.Placeholder),
             # FIXME(andi): VALUES shouldn't be listed here
             # see https://github.com/andialbrecht/sqlparse/pull/64
diff --git a/tests/test_parse.py b/tests/test_parse.py
@@ -2,11 +2,15 @@
 
 """Tests sqlparse function."""
 
+import pytest
+
 from tests.utils import TestCaseBase
 
 import sqlparse
 import sqlparse.sql
 
+from sqlparse import tokens as T
+
 
 class SQLParseTest(TestCaseBase):
     """Tests sqlparse.parse()."""
@@ -139,3 +143,10 @@ def test_psql_quotation_marks():  # issue83
           ....
     $PROC_2$ LANGUAGE plpgsql;""")
     assert len(t) == 2
+
+
+@pytest.mark.parametrize('ph', ['?', ':1', ':foo', '%s', '%(foo)s'])
+def test_placeholder(ph):
+    p = sqlparse.parse(ph)[0].tokens
+    assert len(p) == 1
+    assert p[0].ttype is T.Name.Placeholder
