Pandas UnicodeDecodeError while using read_sql

I am trying to execute SQL queries using pandas.read_sql. It usually works, but for certain queries I run into this error:

  File "C:\Anaconda3\lib\site-packages\pandas\io\sql.py", line 1454, in _fetchall_as_list
    result = cur.fetchall()

    UnicodeDecodeError: 'ascii' codec can't decode byte 0xb4 in position 3: ordinal not in range(128)

I tried solutions suggested for a very similar problem here (UnicodeDecodeError with pandas.read_sql) but it didn't resolve the issue.

I am using cx_oracle library for the database connection.

I tried

db = cx_Oracle.connect(user,pwd, dsn_dict[dbname],encoding='utf-8')

but when I check the encoding using

print(db.encoding)
print(db.nencoding)

I always get

ASCII
ASCII

I tried changing NLS_LANG using

os.environ['NLS_LANG'] = 'AMERICAN_AMERICA.US7ASCII'

but it results in the same error

These are the database NLS parameters:

NLS_CHARACTERSET    US7ASCII

NLS_NCHAR_CHARACTERSET  AL16UTF16

I ran the same query in access and I notice this character in the query result, which might be causing this issue:

¿

Basically, I don't know how to set proper encoding to deal with the problem. Any help is appreciated. Thank you.

SOLUTION:

For reference, I solved this by setting

os.environ['NLS_LANG'] = 'AMERICAN_AMERICA.UTF8'

I dont like doing this though. better solutions are appreciated.