Oracle type NCHAR is not supported by OracleDialect.get_columns

[altvod]

If you try to inspect an Oracle table with NCHAR columns, they are returned as NullType()
The following code demonstrates it:

import sqlalchemy as sa
e = make_engine()  # create your Oracle engine somehow
table = sa.Table(
    'my_table', sa.MetaData(),
    sa.Column(name='my_char', type_=sa.CHAR()),
    sa.Column(name='my_nchar', type_=sa.NCHAR()),
)
table.drop(e, checkfirst=True)
table.create(e)
for col in sa.inspect(e).get_columns('my_table'):
    print(col['name'], repr(col['type']))

It will print:

my_char CHAR(length=1)
my_nchar NullType()

[zzzeek]

OK this just needs to be added to ischema_names and some other places, and by the time I work out all the spots for someones PR, the patch is here so I'll just push that up

[zzzeek]

you can do this for now:

from sqlalchemy.types import NCHAR
from sqlalchemy.dialects.oracle import base
base.ischema_names['NCHAR'] = NCHAR

the reflected datatype will just be missing the length at the moment.

[sqla-tester]

Mike Bayer has proposed a fix for this issue in the master branch:

Add NCHAR to Oracle dialect https://gerrit.sqlalchemy.org/1138

[sqla-tester]

Mike Bayer has proposed a fix for this issue in the rel_1_2 branch:

Add NCHAR to Oracle dialect https://gerrit.sqlalchemy.org/1139