Strip special chars in anonymized bind names

Characters that interfere with "pyformat" or "named" formats in bound
parameters, namely ``%, (, )`` and the space character, as well as a few
other typically undesirable characters, are stripped early for a
:func:`.bindparam` that is using an anonymized name, which is typically
generated automatically from a named column which itself includes these
characters in its name and does not use a ``.key``, so that they do not
interfere either with the SQLAlchemy compiler's use of string formatting or
with the driver-level parsing of the parameter, both of which could be
demonstrated before the fix.  The change only applies to anonymized
parameter names that are generated and consumed internally, not end-user
defined names, so the change should have no impact on any existing code.
Applies in particular to the psycopg2 driver which does not otherwise quote
special parameter names, but also strips leading underscores to suit Oracle
(but not yet leading numbers, as some anon parameters are currently
entirely numeric/underscore based); Oracle in any case continues to quote
parameter names that include special characters.

Fixes: #4837
Change-Id: I21cb654c3e4ef786114160b8b4295242720bf3f9
(cherry picked from commit d7aa017d83b416187b54ad38400475fd86d80671)

diff --git a/doc/build/changelog/unreleased_13/4837.rst b/doc/build/changelog/unreleased_13/4837.rst
new file mode 100644 (file)
index 0000000..def5aaa
--- /dev/null
+++ b/doc/build/changelog/unreleased_13/4837.rst
@@ -0,0 +1,20 @@
+.. change::
+    :tags: bug, sql
+    :tickets: 4837
+
+    Characters that interfere with "pyformat" or "named" formats in bound
+    parameters, namely ``%, (, )`` and the space character, as well as a few
+    other typically undesirable characters, are stripped early for a
+    :func:`.bindparam` that is using an anonymized name, which is typically
+    generated automatically from a named column which itself includes these
+    characters in its name and does not use a ``.key``, so that they do not
+    interfere either with the SQLAlchemy compiler's use of string formatting or
+    with the driver-level parsing of the parameter, both of which could be
+    demonstrated before the fix.  The change only applies to anonymized
+    parameter names that are generated and consumed internally, not end-user
+    defined names, so the change should have no impact on any existing code.
+    Applies in particular to the psycopg2 driver which does not otherwise quote
+    special parameter names, but also strips leading underscores to suit Oracle
+    (but not yet leading numbers, as some anon parameters are currently
+    entirely numeric/underscore based); Oracle in any case continues to quote
+    parameter names that include special characters.

diff --git a/lib/sqlalchemy/sql/elements.py b/lib/sqlalchemy/sql/elements.py
index 2a54529afdbf51e0bebb6c27f65070047108cc0b..3f081de50d4c5b50044c10a8771eed83fac6a938 100644 (file)
--- a/lib/sqlalchemy/sql/elements.py
+++ b/lib/sqlalchemy/sql/elements.py
@@ -1160,7 +1160,13 @@ class BindParameter(ColumnElement):
 
         if unique:
             self.key = _anonymous_label(
-                "%%(%d %s)s" % (id(self), key or "param")
+                "%%(%d %s)s"
+                % (
+                    id(self),
+                    re.sub(r"[%\(\) \$]+", "_", key).strip("_")
+                    if key is not None
+                    else "param",
+                )
             )
         else:
             self.key = key or _anonymous_label("%%(%d param)s" % id(self))

diff --git a/test/sql/test_compiler.py b/test/sql/test_compiler.py
index e56647da408b3af580b4a285cdf38cda1f7ae527..e406a6423a149d5c8abdf49ed33a9e8fd3e6366e 100644 (file)
--- a/test/sql/test_compiler.py
+++ b/test/sql/test_compiler.py
@@ -2970,6 +2970,60 @@ class BindParameterTest(AssertsCompiledSQL, fixtures.TestBase):
         eq_(len(set(pp)), total_params, "%s %s" % (len(set(pp)), len(pp)))
         eq_(len(set(pp.values())), total_params)
 
+    def test_bind_anon_name_no_special_chars(self):
+        for paramstyle in "named", "pyformat":
+            dialect = default.DefaultDialect()
+            dialect.paramstyle = paramstyle
+
+            for name, named, pyformat in [
+                ("%(my name)s", ":my_name_s_1", "%(my_name_s_1)s"),
+                ("myname(foo)", ":myname_foo_1", "%(myname_foo_1)s"),
+                (
+                    "this is a name",
+                    ":this_is_a_name_1",
+                    "%(this_is_a_name_1)s",
+                ),
+                ("_leading_one", ":leading_one_1", "%(leading_one_1)s"),
+                ("3leading_two", ":3leading_two_1", "%(3leading_two_1)s"),
+                ("$leading_three", ":leading_three_1", "%(leading_three_1)s"),
+                ("%(tricky", ":tricky_1", "%(tricky_1)s"),
+                ("5(tricky", ":5_tricky_1", "%(5_tricky_1)s"),
+            ]:
+                t = table("t", column(name, String))
+                expr = t.c[name] == "foo"
+
+                self.assert_compile(
+                    expr,
+                    "t.%s = %s"
+                    % (
+                        dialect.identifier_preparer.quote(name),
+                        named if paramstyle == "named" else pyformat,
+                    ),
+                    dialect=dialect,
+                    checkparams={named[1:]: "foo"},
+                )
+
+    def test_bind_anon_name_special_chars_uniqueify_one(self):
+        # test that the chars are escaped before doing the counter,
+        # otherwise these become the same name and bind params will conflict
+        t = table("t", column("_3foo"), column("4%foo"))
+
+        self.assert_compile(
+            (t.c["_3foo"] == "foo") & (t.c["4%foo"] == "bar"),
+            't._3foo = :3foo_1 AND t."4%foo" = :4_foo_1',
+            checkparams={"3foo_1": "foo", "4_foo_1": "bar"},
+        )
+
+    def test_bind_anon_name_special_chars_uniqueify_two(self):
+
+        t = table("t", column("_3foo"), column("4(foo"))
+
+        self.assert_compile(
+            (t.c["_3foo"] == "foo") & (t.c["4(foo"] == "bar"),
+            't._3foo = :3foo_1 AND t."4(foo" = :4_foo_1',
+            checkparams={"3foo_1": "foo", "4_foo_1": "bar"},
+        )
+
     def test_bind_as_col(self):
         t = table("foo", column("id"))