Very large integers as parameters

[keitherskine]

Environment

To diagnose, we usually need to know the following, including version numbers. On Windows, be
sure to specify 32-bit Python or 64-bit:

• Python: 3.7.5
• pyodbc: 4.0.28
• OS: Windows 10 Pro, 64-bit
• DB: SQL Server 2019 Express
• driver: ODBC Driver 17 for SQL Server

Issue

Python allows integers of more-or-less any size. When inserting a very large integer to a table as a parameter (a value larger than an 8 byte integer can contain), the value of -1 is inserted instead. Oddly, an exception is also raised but not one of the standard PEP249 exceptions. I would expect a PEP249 exception to be raised by pyodbc itself and nothing written to the database. Here is the code that reproduces the issue:

import pyodbc
cnxn = pyodbc.connect(dsn='<my_DSN>', autocommit=True)
crsr = cnxn.cursor()

crsr.execute("DROP TABLE IF EXISTS T1")
crsr.execute("CREATE TABLE T1 (col1 BIGINT NULL)")

try:
    sql = "INSERT INTO T1 (col1) VALUES (?)"
    parameters = (99999999999999999999999999999999999999999999999999,)
    crsr.execute(sql, parameters)
except Exception as err:
    print('exception:', err)

rows = crsr.execute("SELECT * FROM T1").fetchall()
print('table records:', rows)

crsr.close()
cnxn.close()

This code outputs:

exception: <method 'execute' of 'pyodbc.Cursor' objects> returned a result with an error set
table records: [(-1, )]

My guess is that PyLong_AsLongLong is returning -1 (as expected) but then PyErr_Occurred() is not being checked somewhere. Here perhaps?

[v-makouz]

Nice! Looks like a good change, hopefully @mkleehammer can merge it in