[IBMiOSS] Trouble with numeric columns in pyodbc

I am having severe issues with pyodbc in PASE.

Let's say TESTFILE is a table with four columns, TFLD01, TFLD02,
TFLD03, and TFLD04, all numeric. I've tried the following from both my
PC and PASE:

# begin snippet
import pyodbc

conn = pyodbc.connect(<my connection stuff here>)
c1 = conn.cursor()

c1.execute('''
insert into testfile values (?, ?, ?, ?)
''', (1, 2, 3, 4))
c1.execute('''
insert into testfile values (?, ?, ?, ?)
''', ('2', '4', '6', '8'))
# end snippet

From my PC, both rows are successfully inserted. The second row is

implicitly converted to numeric values.

From PASE, it's a disaster. The first row is silently inserted as all

zeros (the columns are not null-capable; had they been, I suppose the
row might have consisted of nulls), and the second row raises the
following exception:

pyodbc.DataError: ('22018', '[22018] [IBM][System i Access ODBC
Driver][DB2 for i5/OS]SQL0406 - Conversion error on assignment to
column TFLD01. (-406) (SQLExecDirectW)')

Is anyone else experiencing this? In its current state, at least on my
machine, obviously pyodbc is unusable from PASE. :(

John Y.