psycopg2 selecting timestamp returns datetime.datetime wrapped in tuple, how to unpack?

Question

I have created a table:

 cursor.execute("CREATE TABLE articles (title varchar PRIMARY KEY, pubDate timestamp with time zone);")

I inserted a timestamp like this:

timestamp = date_datetime.strftime("%Y-%m-%d %H:%M:%S+00")

cursor.execute("INSERT INTO articles VALUES (%s, %s)", 
              (title, timestamp))

When I run a SELECT statement to retrieve the timestamps, it returns tuples instead:

cursor.execute("SELECT pubDate FROM articles")
rows = cursor.fetchall()
for row in rows:
    print(row)

This is the returned row:

(datetime.datetime(2015, 12, 9, 6, 47, 4, tzinfo=psycopg2.tz.FixedOffsetTimezone(offset=660, name=None)),)

How can I retrieve the datetime object directly?

I've looked up a few other related questions (see here and here) but can't seem to find the answer. Probably overlooking something simple here but any help would be much appreciated!

Answer 1

Python's datetime objects are automatically adapted into SQL by psycopg2, you don't need to stringify them:

cursor.execute("INSERT INTO articles VALUES (%s, %s)", 
               (title, datetime_obj))

To read the rows returned by a SELECT you can use the cursor as an iterator, unpacking the row tuples as needed:

cursor.execute("SELECT pubDate FROM articles")

for pub_date, in cursor:  # note the comma after `pub_date`
    print(pub_date)

Answer 2

After some more googling I think I figured it out. If I change:

print(row)

to

print(row[0])

It actually works. I guess this is because row is a tuple and this is way to unpack the tuple correctly.