Help needed : Dictionary tricks in 2.2

Alex Martelli aleax at
Mon Jan 14 14:46:38 CET 2002

"pekka niiranen" <krissepu at> wrote in message
news:3C429CF5.60700 at
> 2) The following code turns list (or tuple) into dictionary fast (without
> >>> d = {}
> >>> l = ['a','b','c','b']
> >>> dummy = map(operator.setitem, [d]*len(l), l, [])

Ugly.  Marginally better in 2.2
    map(d.__setitem__, l, [])
but it's still pretty obscure.

> >>> d
> {'a': None, 'c': None, 'b': None}
> Q2: How can I modify the code to create dictionary with
>     empty string -values instead of None's
> {'a': "", 'c': "", 'b': ""}

map(d.__setitem__, l, len(l)*[''])

but it's STILL a hack.  Much clearer and more readable:

    for item in l: d[item] = ''

and you'll likely never notice the performance-loss -- this
isn't going to be your application's bottleneck!

"Premature optimization is the root of all evil" (Knuth).

Target *clarity* first -- simplicity, readability, clean
and straightforward code.  Time to squeeze out the last
drop of performance, *IF EVER*, comes once your program
is working, simple, and clean, and ONLY IF it's not fast
enough, and ONLY on those tiny areas of code that are in
fact part of a program's performance "bottleneck" (find
out by profiling -- intuition is misleading here).


More information about the Python-list mailing list