The Pythonista's Legacy

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

#!/usr/bin/env python3
#
# So, I've got some free python training for you! This code works both
# with python2 and python3. I love python3's print function, so let's add
# it...

from __future__ import print_function

# Ok. First thing, did you know about namedtuple? It works as a regular
# tuple, but you can access the fields by name (hence "named" tuple).
# Besides, it is optimized for producing a large amount of instances.

from collections import namedtuple

# It works as a meta-type: you define types by calling a function.
Row = namedtuple("Row", "name age city")

# ...but refer to the documentation (pydoc collections.namedtuple) for
# that. Here a proof of how it works:

r = Row("Luke", 40, "Amsterdam")
print("Showing Row:", r)
print("Showing fields of the row:", r.name, r.age, r.city)

# Let's take an hypotetical list of results. This Could be the output of
# some stats file, or maybe from a database (even if the database will
# probably have some inner type!).

records = [
    Row("Jane", 15, "London"),
    Row("Georg", 15, "Berlin"),
    Row("Urlika", 20, "Stockholm"),
    Row("Johan", 17, "Stockholm"),
    Row("Aldo", 9, "Rome"),
    Row("Hans", 17, "Berlin")
    Row("Petra", 21, "Berlin")
]

# About itertools? It's an extremely nice library of python, with useful
# lazy-evaluated functions [ https://en.wikipedia.org/wiki/Lazy_evaluation ].
# Iterators in general allow to iterate (hence the word) on objects like
# tuples or lists ("iterables").

import itertools as it

# Iterators are very elegant, and if wisely used they can reduce the
# memory footprint of a program, but on the minus side they get consumed:
# you cannot use the same iterator twice. Example:

# I can print lists twice:

print('-' * 80)
print("Records")
for r in records:
    print("\t", r)
print("...second shot:")
for r in records:
    print("\t", r)
print("end.")

print('-' * 80)

# But iterators get consumed
records_iterator = iter(records)
print("Iterated records")
for r in records_iterator:
    print("\t", r)
print("...second shot (will be empty):")
for r in records_iterator:
    print("\t", r)
print("end.")
print('-' * 80)

# Also I'll include the "attrgetter" operator. Quoting the documentation:
#
#   After, f=attrgetter('name'), the call f(r) returns r.name.

from operator import attrgetter as aget

# For example, this is get_city:
get_city = aget('city')

# Since the records are namedtuple with a field called `city`,
# `aget('city')` will produce a function returning the `city` field:
r = Row("Luke", 40, "Amsterdam")
print("The city of Luke is:", get_city(r))

# Now some interesting use of the groupby function.
#
# It basically works as the "Group By" operator on SQL: group together
# items of a table (in this case an iterator), and allow iteration over
# single groups. The only caveat: items of the same group must be
# contiguous, as for the "uniq" command of unix, which is often preceded
# by the "sort" command, in pipe.
#
# For instance, let's take the `records` list, we defined previously, and
# group the rows by the `city` attribute.  The `get_city` we defined
# before turns out to be a good grouping function So, those items have to
# be sorted by the relevant field...
records_sorted = sorted(records, key=get_city)

# ...Then we can use the groupby:
for city, rows in it.groupby(records_sorted, get_city):
    print("City:", city, end="\n\t")
    print(*rows, sep="\n\t", end="\n\n")

# Nice thing: this allows us to do some aggregation (e.g. sum, or average)
# quite easily. Want to know the average age by city?
for city, rows in it.groupby(records_sorted, get_city):
    print(
        "City:", city,
        "Avg.age:", sum(map(aget('age'), rows)),
        end="\n\n"
    )

# Note: sorted() will return *a list*, not an iterator. That's why we can
# use the it.groupby() function twice on it. If line 99 were
#
#   records_sorted = iter(sorted(records, key=get_city))
#
# Then the second cycle at line 108 would have not worked. Also you may
# try to cycle with it.groupby on records instead of records_sorted: you
# will see the results! :)

# Happy hacking.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=+CMW
-----END PGP SIGNATURE-----