Comments on: CodeKata 4: Data Munging

By: Weekly Digest for January 3rd | William Stearns

Weekly Digest for January 3rd | William Stearns — Mon, 04 Jan 2010 04:24:35 +0000

[...] Shared Brian Jones: CodeKata 4: Data Munging. [...]

By: m0j0

m0j0 — Tue, 29 Dec 2009 21:40:23 +0000

Tests? Care to clarify?

By: Herberth Ama!ral

Herberth Ama!ral — Tue, 29 Dec 2009 12:59:34 +0000

Cool. But…. where is the tests?!

By: m0j0

m0j0 — Tue, 29 Dec 2009 12:51:02 +0000

That last bit about using min() on the iterator stream is cool – thanks for posting that. The ‘line[:1] in string.digits’… I’d still have to strip it — I believe both data sets have lines starting with tabs/whitespace. Other than that it’s a good idea — I think your way is more readable.

Using a function for the fields would work, and in a larger piece of code it would be a real no-brainer, but I think there’s a threshold where if the code is under a certain number of lines or something, then adding function calls makes things more complex than is necessary. I guess it’s an academic exercise, though, so…. you’re right

By: Andrew Dalke

Andrew Dalke — Tue, 29 Dec 2009 07:30:40 +0000

Here’s a few changes you might try. Replace “line.strip().startswith(tuple(string.digits)) with “line[:1] in string.digits”

Use a function to get the fields for each of the “football” and “weather” cases, then pull the “if ‘….’ in sys.argv[1]:” out of the loop and instead use the appropriate field extraction function.

Even more unusual, convert the code to a function which returns each diff and value through an iterator, then use min() on the iterator stream from the generator. It isn’t cleaner or shorter than what you have (I think it’s longer and more complex, actually), but it’s an interesting variant.