On 2013-12-08 15:04, Peter Otten wrote:
> > data = dict(
> > (str(date), {source: freq})
> > for date, freq, source in original_data
> > )
>
> or even just
>
> data = {str(date): {source: freq}
> for date, freq, source in original_data}
I maintain enough pre-2.7 code that I tend to eschew
dict-comprehensions for the time being. I like them as a language
addition, but can't use them yet in most of the code-bases with which
I work. To the OP, if you don't have to support pre-2.7 code, then
this is a tidier way to go.
> But do you really need a dict with a single key? And is it even
> correct? If a date occurs twice only the last source:freq pair is
> kept. Without knowing the context the humble
>
> data = {}
> for date, freq, source in original_data:
> source_to_freq = data.setdefault(date, {})
> if source in source_to_freq:
> raise ValueError(
> "Multiple frequencies for one source not supported")
> source_to_freq[source] = freq
>
> appears so much more plausible...
The OP's code didn't do any such sanity-checking so I made the
assumption that the data came in clean. I know better than to trust
incoming data (even my own) in my own code, but...
(followup to the OP's reply elsewhere in the thread coming shortly)
-tkc
--
https://mail.python.org/mailman/listinfo/python-list
