Dear counters,
I have also encountered the above problem, and I frequently use the two
blocks of code below.
The first one is commented already - except that this version does not
start from an empty dict.
The second "inverts" the first one. Meaning that a dict in word2hits
style:
my_dict['the'] = 120
my_dict['word'] = 4
...
my_dict['gene'] = 4
becomes a dict in hits2words style:
new_dict[4] = ['word','gene', ...]
new_dict[120] = ['the', ...]
I used these to count genes in in abstracts from pubmed the when I
"invented" them (but it took some time to remove them from the code and
use them as functions).
def dict_add(indict,inlist,init=1):
for item in inlist:
if indict.has_key(item):
indict[item] += 1
else:
indict[item] = init
return indict
# end def
def dict_invert(indict):
new_dict = {}
for key in indict.keys():
if new_dict.has_key(indict[key]):
new_dict[indict[key]].append(key)
else:
new_dict[indict[key]] = [key]
return new_dict
#end def
A good idea could be to change the header of the first one (if you want
the option to start counting from zero) into:
def dict_add(inlist,indict={},init=1):
/P9K
--
http://mail.python.org/mailman/listinfo/python-list
