OK thanks for looking into this and I'm glad we found the issue,
mostly thanks to your prompt and accurate replies so thanks!
On Wed, Jul 5, 2017 at 9:46 AM, Paul Morel
wrote:
> It is possible that in the end that was the case, however, the reason I even
>
It is possible that in the end that was the case, however, the reason I
even attempted the pip uninstall/install originally was because of the
error I was getting. I found some SO posts that recommended rolling back
to an earlier version of SQLAlchemy. Rolling back had no impact. I then
On Jul 5, 2017 12:59 AM, "Paul Morel"
wrote:
It appears that the Anaconda installer placed the .pyd files there. I
discovered that pip uninstall doesn't remove those files like I had
assumed. Therefore, when I pip uninstalled/installed the original .pyd
files
It appears that the Anaconda installer placed the .pyd files there. I
discovered that pip uninstall doesn't remove those files like I had
assumed. Therefore, when I pip uninstalled/installed the original .pyd
files were still there.
It looks like deleting the .pyd files and pip installing
Is there a chance you updated the Python binary as part of your process?
Weird stuff like this often happens if you forget to trash the
site-packages or virtualenv after a python version upgrade (even on a micro
version upgrade)
--
SQLAlchemy -
The Python SQL Toolkit and Object Relational
Ok how did the pyd file get there in the first place, was that pre-built as
part of anaconda or did it compile on your local machine? This is kind of
a serious issue that a broken shared library got created
On Jul 4, 2017 7:55 PM, "Paul Morel" wrote:
BAM! That
BAM! That was it. There must be some type of incompatibility between the
c extensions and the Windows 2008 Server configuration. I pip
uninstalled/installed and found the c extensions where rebuilt. After
removing *cresultproxy.pyd* from the sqlalchemy site-packages directory the
queries
it's the select schema_name() query that's failing and that
cursor.description is fine.
did you build the SQLAlchemy C extensions on windows? or does your
anaconda build somehow deliver sqlalchemy/cresultproxy.dll to your
installation ?my last theory here is that you have C extensions
set
Sorry, I gave you the cursor.description of the schema_name() query. This
is the output for the cursor.description for the SELECT * FROM EPO_MODELS
query:
*pymssql Cursor Description is:*
> ((u'ID', 3, None, None, None, None, None), (u'MODELTYPE', 3, None, None,
> None, N
> one, None),
>
> Testing *pymssql* Directly
>
> Complete
> Testing Get Schema Name with pymssql
> (u'dbo',)
>
> *Cursor Description is:((u'', 1, None, None, None, None, None),)*
> Complete
> Testing *pyodbc* Directly
>
> Complete
> Testing Get Schema Name
in your direct examples can you show me the output of
"cursor.description" as well once you execute the query?
On Tue, Jul 4, 2017 at 11:17 AM, Paul Morel
wrote:
> I ran that select statement under both the pyodbc and pymssql direct
> connections and it appears to
I ran that select statement under both the pyodbc and pymssql direct
connections and it appears to have returned a legitimate row of data. This
is the excerpt from the test output:
Testing pymssql Directly
>
> Complete
> Testing Get Schema Name with pymssql
> (u'dbo',)
>
It's never a good idea to just blanket ignore all exceptions. There
are a few cases where we except for dbapi.Error in order to catch
unsupported calls such as looking for sys.dm_exec_sessions.
However here, SQL Server is not complaining at all. A result is
returned. The result has a row.
can you run this query please?
SELECT schema_name()
the issue is, that query is returning a result, there is a row, but it
no columns, which is nonsensical. Did you try running with the
mssql+pymssql:// driver? Looks like a pyodbc bug so far but need
more info.
On Tue, Jul 4, 2017 at 9:59
I don't know the downstream implications of doing this but would it make
sense to wrap lines 1773-1777 in a try/except with the except block
returning self.schema_name.
Like this:
def _get_default_schema_name(self, connection):
if self.server_version_info < MS_2005_VERSION:
return
The key part of the stack trace is:
File "c:\Program
Files\Anaconda2\lib\site-packages\sqlalchemy\dialects\mssql\base.py",
line 1773, in _get_default_schema_name default_schema_name =
connection.scalar(query)
...which is in this function:
Mike,
Sorry for the lack of information. Please find the rest of what you wanted
below.
*Full Stack Trace:*
Running Direct SQL Query
> Traceback (most recent call last):
> File "test.py", line 45, in result = con.execute('SELECT * FROM
> EPO_MODELS')
> File "c:\Program
>
Can you send complete stack trace please? That's the main thing that will
show which query this is occurring on (there are several upon connect).
Also full detail on SQL server version, odbc driver, client operating
system.
On Jul 3, 2017 7:26 PM, "Paul Morel"
18 matches
Mail list logo