- Support has been improved for Postgresql reflection behavior on very old

(pre 8.1) versions of Postgresql, and potentially other PG engines
such as Redshift (assuming Redshift reports the version as < 8.1).
The query for "indexes" as well as "primary keys" relies upon inspecting
a so-called "int2vector" datatype, which refuses to coerce to an array
prior to 8.1 causing failures regarding the "ANY()" operator used
in the query.  Extensive googling has located the very hacky, but
recommended-by-PG-core-developer query to use when PG version < 8.1
is in use, so index and primary key constraint reflection now work
on these versions.

6 files changed, 55 insertions, 31 deletions

diff --git a/doc/build/changelog/changelog_09.rst b/doc/build/changelog/changelog_09.rst
index caf5ced07..f8a9c173e 100644
--- a/doc/build/changelog/changelog_09.rst
+++ b/doc/build/changelog/changelog_09.rst
@@ -15,6 +15,20 @@
     :version: 0.9.3
 
     .. change::
+        :tags: postgresql, bug
+
+        Support has been improved for Postgresql reflection behavior on very old
+        (pre 8.1) versions of Postgresql, and potentially other PG engines
+        such as Redshift (assuming Redshift reports the version as < 8.1).
+        The query for "indexes" as well as "primary keys" relies upon inspecting
+        a so-called "int2vector" datatype, which refuses to coerce to an array
+        prior to 8.1 causing failures regarding the "ANY()" operator used
+        in the query.  Extensive googling has located the very hacky, but
+        recommended-by-PG-core-developer query to use when PG version < 8.1
+        is in use, so index and primary key constraint reflection now work
+        on these versions.
+
+    .. change::
         :tags: sqlite, bug
 
         The SQLite dialect will now skip unsupported arguments when reflecting
@@ -58,8 +72,8 @@
 
         Added server version detection to the newly added dialect startup
         query for  "show standard_conforming_strings"; as this variable was
-        added as of PG 8.2, we skip the query for PG versions older than
-        that as well as for backends like Redshift.
+        added as of PG 8.2, we skip the query for PG versions who report a
+        version string earlier than that.
 
     .. change::
         :tags: bug, orm, declarative
diff --git a/lib/sqlalchemy/dialects/postgresql/base.py b/lib/sqlalchemy/dialects/postgresql/base.py
index 7b9ee799e..cea9d67b6 100644
--- a/lib/sqlalchemy/dialects/postgresql/base.py
+++ b/lib/sqlalchemy/dialects/postgresql/base.py
@@ -2031,35 +2031,22 @@ class PGDialect(default.DefaultDialect):
         table_oid = self.get_table_oid(connection, table_name, schema,
                                        info_cache=kw.get('info_cache'))
 
-        if self.server_version_info < (8, 0):
-            # the shortcoming of this query is that it will
-            # not detect a PK constraint that has been renamed.
-            # This query was removed with #2291, however it was reported
-            # that the newer queries do not work with PG 7 so here
-            # it is restored when old PG versions are detected.
-            PK_SQL = """
-              SELECT attname FROM pg_attribute
-              WHERE attrelid = (
-                 SELECT indexrelid FROM pg_index i
-                 WHERE i.indrelid = :table_oid
-                 AND i.indisprimary = 't')
-              ORDER BY attnum
-            """
-        elif self.server_version_info < (8, 4):
-            # unnest() and generate_subscripts() both introduced in
-            # version 8.4
+        if self.server_version_info < (8, 4):
             PK_SQL = """
                 SELECT a.attname
                 FROM
                     pg_class t
                     join pg_index ix on t.oid = ix.indrelid
                     join pg_attribute a
-                        on t.oid=a.attrelid and a.attnum=ANY(ix.indkey)
+                        on t.oid=a.attrelid AND %s
                  WHERE
                   t.oid = :table_oid and ix.indisprimary = 't'
                 ORDER BY a.attnum
-            """
+            """ % self._pg_index_any("a.attnum", "ix.indkey")
+
         else:
+            # unnest() and generate_subscripts() both introduced in
+            # version 8.4
             PK_SQL = """
                 SELECT a.attname
                 FROM pg_attribute a JOIN (
@@ -2174,6 +2161,21 @@ class PGDialect(default.DefaultDialect):
             fkeys.append(fkey_d)
         return fkeys
 
+    def _pg_index_any(self, col, compare_to):
+        if self.server_version_info < (8, 1):
+            # http://www.postgresql.org/message-id/10279.1124395722@sss.pgh.pa.us
+            # "In CVS tip you could replace this with "attnum = ANY (indkey)".
+            # Unfortunately, most array support doesn't work on int2vector in
+            # pre-8.1 releases, so I think you're kinda stuck with the above
+            # for now.
+            # regards, tom lane"
+            return "(%s)" % " OR ".join(
+                        "%s[%d] = %s" % (compare_to, ind, col)
+                        for ind in range(0, 10)
+                    )
+        else:
+            return "%s = ANY(%s)" % (col, compare_to)
+
     @reflection.cache
     def get_indexes(self, connection, table_name, schema, **kw):
         table_oid = self.get_table_oid(connection, table_name, schema,
@@ -2186,14 +2188,14 @@ class PGDialect(default.DefaultDialect):
           SELECT
               i.relname as relname,
               ix.indisunique, ix.indexprs, ix.indpred,
-              a.attname, a.attnum, ix.indkey::varchar
+              a.attname, a.attnum, ix.indkey%s
           FROM
               pg_class t
                     join pg_index ix on t.oid = ix.indrelid
                     join pg_class i on i.oid=ix.indexrelid
                     left outer join
                         pg_attribute a
-                        on t.oid=a.attrelid and a.attnum=ANY(ix.indkey)
+                        on t.oid=a.attrelid and %s
           WHERE
               t.relkind = 'r'
               and t.oid = :table_oid
@@ -2201,7 +2203,10 @@ class PGDialect(default.DefaultDialect):
           ORDER BY
               t.relname,
               i.relname
-        """
+        """ % (
+                "::varchar" if self.server_version_info >= (8, 1) else "",
+                self._pg_index_any("a.attnum", "ix.indkey")
+            )
 
         t = sql.text(IDX_SQL, typemap={'attname': sqltypes.Unicode})
         c = connection.execute(t, table_oid=table_oid)
diff --git a/test/dialect/postgresql/test_dialect.py b/test/dialect/postgresql/test_dialect.py
index fd6df2c98..7f22c6a6b 100644
--- a/test/dialect/postgresql/test_dialect.py
+++ b/test/dialect/postgresql/test_dialect.py
@@ -205,6 +205,7 @@ class MiscTest(fixtures.TestBase, AssertsExecutionResults, AssertsCompiledSQL):
                     : Numeric})
         assert_raises(exc.InvalidRequestError, testing.db.execute, stmt)
 
+    @testing.only_if("postgresql >= 8.2", "requires standard_conforming_strings")
     def test_serial_integer(self):
 
         for version, type_, expected in [
diff --git a/test/dialect/postgresql/test_query.py b/test/dialect/postgresql/test_query.py
index ee5800db0..82b49d974 100644
--- a/test/dialect/postgresql/test_query.py
+++ b/test/dialect/postgresql/test_query.py
@@ -70,6 +70,7 @@ class InsertTest(fixtures.TestBase, AssertsExecutionResults):
         metadata.create_all()
         self._assert_data_with_sequence(table, 'my_seq')
 
+    @testing.requires.returning
     def test_sequence_returning_insert(self):
         table = Table('testtable', metadata, Column('id', Integer,
                       Sequence('my_seq'), primary_key=True),
@@ -84,6 +85,7 @@ class InsertTest(fixtures.TestBase, AssertsExecutionResults):
         metadata.create_all()
         self._assert_data_autoincrement(table)
 
+    @testing.requires.returning
     def test_opt_sequence_returning_insert(self):
         table = Table('testtable', metadata, Column('id', Integer,
                       Sequence('my_seq', optional=True),
@@ -97,6 +99,7 @@ class InsertTest(fixtures.TestBase, AssertsExecutionResults):
         metadata.create_all()
         self._assert_data_autoincrement(table)
 
+    @testing.requires.returning
     def test_autoincrement_returning_insert(self):
         table = Table('testtable', metadata, Column('id', Integer,
                       primary_key=True), Column('data', String(30)))
@@ -592,8 +595,7 @@ class ServerSideCursorsTest(fixtures.TestBase, AssertsExecutionResults):
 
 class MatchTest(fixtures.TestBase, AssertsCompiledSQL):
 
-    __only_on__ = 'postgresql'
-    __excluded_on__ = ('postgresql', '<', (8, 3, 0)),
+    __only_on__ = 'postgresql >= 8.3'
 
     @classmethod
     def setup_class(cls):
diff --git a/test/dialect/postgresql/test_reflection.py b/test/dialect/postgresql/test_reflection.py
index 705a64c8e..aefb6a0ba 100644
--- a/test/dialect/postgresql/test_reflection.py
+++ b/test/dialect/postgresql/test_reflection.py
@@ -20,7 +20,7 @@ class DomainReflectionTest(fixtures.TestBase, AssertsExecutionResults):
 
     """Test PostgreSQL domains"""
 
-    __only_on__ = 'postgresql'
+    __only_on__ = 'postgresql > 8.2'
 
     @classmethod
     def setup_class(cls):
@@ -128,8 +128,8 @@ class DomainReflectionTest(fixtures.TestBase, AssertsExecutionResults):
 class ReflectionTest(fixtures.TestBase):
     __only_on__ = 'postgresql'
 
-    @testing.fails_if(('postgresql', '<', (8, 4)),
-            "newer query is bypassed due to unsupported SQL functions")
+    @testing.fails_if("postgresql < 8.4",
+                    "Better int2vector functions not available")
     @testing.provide_metadata
     def test_reflected_primary_key_order(self):
         meta1 = self.metadata
@@ -170,6 +170,8 @@ class ReflectionTest(fixtures.TestBase):
         eq_(
             t.c.x.server_default.arg.text, "'%s'::character varying" % ("abcd" * 40)
         )
+
+    @testing.fails_if("postgresql < 8.1", "schema name leaks in, not sure")
     @testing.provide_metadata
     def test_renamed_sequence_reflection(self):
         metadata = self.metadata
diff --git a/test/dialect/postgresql/test_types.py b/test/dialect/postgresql/test_types.py
index ba4b63e1a..4540b7537 100644
--- a/test/dialect/postgresql/test_types.py
+++ b/test/dialect/postgresql/test_types.py
@@ -102,7 +102,7 @@ class FloatCoercionTest(fixtures.TablesTest, AssertsExecutionResults):
 
 class EnumTest(fixtures.TestBase, AssertsExecutionResults):
 
-    __only_on__ = 'postgresql'
+    __only_on__ = 'postgresql > 8.3'
 
 
     @testing.fails_on('postgresql+zxjdbc',
@@ -908,7 +908,7 @@ class SpecialTypesTest(fixtures.TestBase, ComparesTables, AssertsCompiledSQL):
 class UUIDTest(fixtures.TestBase):
     """Test the bind/return values of the UUID type."""
 
-    __only_on__ = 'postgresql'
+    __only_on__ = 'postgresql >= 8.3'
 
     @testing.fails_on('postgresql+zxjdbc',
                       'column "data" is of type uuid but expression is of type character varying')