Adding "Streaming" cursor functionality

[gene-peters]

When pulling result rows from Thrift, the current cursor attempts to load them all into a single variable which it then returns. This can cause major issues with large result sets; I've implemented a cursor class which returns a generator object and iterates over each result row.

Example usage:

from cursor import SSCursor

conn = pyhs2.connect(cursorclass=SSCursor, **kwargs)
cur = conn.cursor()
cur.execute("SELECT * FROM table LIMIT 100")
for result in cur.fetch():
...print result

[davidbgk]

I second that PR, this is very important when you deal with big data (otherwise your ram will explode).

[ChuckConnell]

Please see this long discussion #8 on the same topic.