[Sqlalchemy-tickets] Issue #3586: numpy int64 garbled upon insert (zzzeek/sqlalchemy)

[Tim D. <iss...@bi...>]

New issue 3586: numpy int64 garbled upon insert
https://bitbucket.org/zzzeek/sqlalchemy/issues/3586/numpy-int64-garbled-upon-insert

Tim Diels:

`numpy.int64` is garbled upon insert. This may apply to other numpy types as well. Workaround is to cast with `int` before inserting.

Problem observed on latest commit (0847097c29ab48a5f111518e2c6ee324d5242057) and on v1.0.8.

Test case:

```
#!python

    from sqlalchemy import create_engine, MetaData, Column, Integer, Table
    import numpy as np
    engine = create_engine('sqlite:///test.db', echo=True)
    metadata = MetaData()
    test_table = Table('test', metadata,
        Column('id', Integer, primary_key=True),
        Column('number', Integer)
    )
    metadata.drop_all(engine)  # clean
    metadata.create_all(engine)
    engine.execute(test_table.insert(), {'number': np.int64(123)})
    result = engine.execute(test_table.select()).first()
    assert result[1] == 123, result
```