psycopg2: DictCursor vs RealDictCursor

The main advantage of real dictionary cursor is the easiness to get a query output as json.

Compare:

with psycopg2.connect('dbname=test') as connection:    with connection.cursor(cursor_factory=RealDictCursor) as cursor:        cursor.execute("select * from my_table")        print(json.dumps(cursor.fetchall()))

versus

with psycopg2.connect('dbname=test') as connection:    with connection.cursor() as cursor:        cursor.execute("select * from my_table")        columns = [desc[0] for desc in cursor.description]        real_dict = [dict(zip(columns, row)) for row in cursor.fetchall()]        print(json.dumps(real_dict))

There is no important difference between these options when it comes to performance.

You cannot get an expected json using json.dumps(cursor.fetchall()) for regular or dictionary-like cursors and need the conversion showed above. On the other hand, real dictionary cursor produces a much larger result so you should not use it if you really do not need it.

class psycopg2.extras.RealDictCursor(*args, **kwargs)

A cursor that uses a real dict as the base type for rows. Note that this cursor is extremely specialized and does not allow the normal access (using integer indices) to fetched data. If you need to access database rows both as a dictionary and a list, then use the generic DictCursor instead of RealDictCursor. class psycopg2.extras.RealDictConnection A connection that uses RealDictCursor automatically.

NoteNot very useful since Psycopg2.5: you can use psycopg2.connect(dsn, cursor_factory=RealDictCursor) instead of RealDictConnection. class psycopg2.extras.RealDictRow(cursor) A dict subclass representing a data record.