Review of a Python Video

Several days ago, I watched a video called Transforming Code into Beautiful, Idiomatic Python on twitter, tweeted by @pythoncn. This video was given by @raymondh, who is a Python core developer and a very attractive programmer and teacher. This review almost covers everything I learnt from the video. Thank you, Mr. Hettinger.

Notes

Notification: everything listed here is based on Python 2.7, something may be a little different in Python 3, for example, range in Python 3 is actually xrange in Python 2.7

• xrange is far more better than range, izip is better than zip, because iterator is a good thing!
• Use key in sorted() instead of cmp, because it’s sufficient and effecient.
• A neat way to call a function until a sentinel value.

For example, here is a problem, we want seperate the file f into a list of 32 bytes long blocks. The common code would be like:

    blocks = []
    while True:
        block = f.read(32)
        if block == '':
            break
        blocks.append(block)

Another way of doing this kind of repeating-functions-with-sentinel-value could exactly be done like:

    blocks = []
    for block in iter(partial(f.read, 32), '')
        blocks.append(block)

iter(function, sentinel_value) keeps calling the function until it returns the sentinel value. partial(function, *arg) function passes the argument list to the function. The second solution is more neat and beautiful.

• Difference between two ways of iterating a dictionay.

    d = {'matthew': 'blue', 'rachel': 'green', 'raymond': 'red'}
    # Just iterating, no mutable action
    for k in d:
        print k
    # Keys can be deleted in this way, for d.keys() is another list of keys.
    for k in d.keys():
        if k.startswith('r'):
            del d[k]

Otherwise, interpreter will raise a Runtime Error: dictionary changed size during iteration.

• iteritems() is better than items()
• Use defaultdict().
• Clarify function calls with keyword arguments.
• namedtuple is a nice tool.
• Updating multiple state variables simultaneously.
• Use deque for efficiency.
• Use with clause.

    with open('file', 'r') as f:
        d = f.read()

These are basicly everything I learned from the video, again, thank you, Mr. Hettinger.