itertools.groupby

Carsten Haese carsten at uniqsys.com
Mon May 28 09:17:55 EDT 2007


On Sun, 2007-05-27 at 18:12 -0700, Steve Howell wrote:
> [...] there is no way
> that "uniquekeys" is a sensible variable [...]

That's because the OP didn't heed the advice from the docs that
"Generally, the iterable needs to already be sorted on the same key
function."

> http://informixdb.blogspot.com/2007/04/power-of-generators-part-two.html
> > 
> 
> Do we now, or could we, link to this example from the
> docs?

Do we? No. Could we? Technically we could, but I'm not sure we should.
The article is about more than just groupby. I just brought it up as one
readily available instance of an independent source of examples.

> To repeat myself, I think a concrete example is
> beneficial even on the main page:

I disagree. Examples add clutter to the page of synopses. Somebody
looking for examples should look on the page that's conveniently called
Examples.

Suppose hypothetically you wanted to show off a really neat example that
involves chain, izip, and groupby. If the examples were forced into the
page of function synopses, you'd have to duplicate it in all three
functions, or randomly pick one function for which your example is an
example. Having a separate examples page that is not arbitrarily
sectioned by function name makes more sense.

Best regards,

-- 
Carsten Haese
http://informixdb.sourceforge.net





More information about the Python-list mailing list