Let's say I have a list
l = ['michael','michael','alice','carter']
I want to map it to the following:
k = [1,1,2,3]
Where michael corresponds to 1, alice corresponds to 2 etc. Is there a function in Python to do this easily?
Add a comment
|
10 Answers
Have a look at ord, which gives the unicode number for a given character:
>>> letters = ['a','b','c','d','e','f','g']
>>> [ord(x) for x in letters]
[97, 98, 99, 100, 101, 102, 103]
So you could do ord(x)-96 to convert a-z to 1-26 (careful about upper case, etc).
l = ['a','b','a','c']
k = [ord(x)-96 for x in l] # [1,2,1,3]
Again, careful about upper case and non-alphabet characters.
answered Feb 9, 2012 at 7:00
mathematical.coffee
57.1k15 gold badges160 silver badges197 bronze badges
3 Comments
user2739472
This only works for a list of characters, not for a list of strings as in the OP's question.
mathematical.coffee
Note that the original question only had a list of characters. The edit to the question was made after this answer was given and accepted!
user2739472
Ah I didn't notice that. In that case, I wish the OP had asked the 'string' question as a separate question. Thanks.
How about using Pandas?
import pandas as pd
l = ['michael','michael','alice','carter']
pd.Series(l).astype('category').cat.codes.values
Comments
In order to answer the edited question, i.e., to map the list of strings to unique integers, one has to first find the unique strings and then do 1-1 mapping of the strings to integers in the original list of strings. For example,
s = ['michael','michael','alice','carter']
then unique strings are {'michael','alice','carter'}. Now, convert these strings to integers by 1-1 mapping like {'michael','alice','carter'} =[1,2,3] using dictionary {'michael':1,'alice':2,'carter':3}. In the third step, loop through the original list of strings; search the string in the dictionary for the corresponding integer and replace the string by that integer.
s=['michael','michael','alice','carter']
mydict={}
i = 0
for item in s:
if(i>0 and item in mydict):
continue
else:
i = i+1
mydict[item] = i
k=[]
for item in s:
k.append(mydict[item])
Output:
k=[1, 1, 2, 3]
Comments
To map list of integers to list of strings I would use a dictionary, for example:
> name_number = {'michael':1, 'michael':1, 'alice':2, 'carter':3}
> print len(name_number)
3
> print name_number['alice']
2
Note that len(name_number) is 3, because duplicate keys are not allowed.
Comments
If I'm reading you correctly, you want to take a list of characters and convert them to integers, with a being 1, b being 2, etc.
l = ['a','b','a','c']
k = [ord(x.upper()) - 64 for x in l]
Threw the upper() in there so it doesn't matter whether they're upper case or lower.
Comments
The function is zip
E.g:
>>> l = ['a','b','a','c']
>>> k = [1,2,1,3]¨
>>> zip(l,k)
[('a', 1), ('b', 2), ('a', 1), ('c', 3)]
If you want to use the items of l as index, you want an dictionary:
>>> d = dict(zip(l,k))
>>> d
{'a': 1, 'c': 3, 'b': 2}
>>> d['a']
1
>>> d['c']
3
>>>
1 Comment
vikki
zip does not check for correspondence, if k is reversed, the results will be different.
can do it pretty easy without a function :
j - list()
for i in range (len(l)) :
j.append((l[i],k[i]))
Comments
From your question it is not clear if you want to generate k based on l or both l and k are given.
If you are looking to create k based on l, @mathematical.coffee's answer should do.
If you want a map from items in l to k, obviously, your items in l should be unique.
See if this is what you were looking for
dict((l[index], k[index]) for index in range(len(l)))
Or else, if you are looking for tuples:
[(l[index], k[index]) for index in range(len(l))]
Comments
If you don't care about the order of the assigned ids, this works:
# create unique list of names
unique_l = set(l)
# create mappings from names to id
name2id = {name: idx+1 for idx, name in enumerate(unique_l)}
# map initial list of names to ids
k = [name2id[name] for name in l]
Output:
[2, 2, 1, 3]
Comments
If you do not mind missing the mapping restriction, probably this is another good may to do it.
import numpy as np
names, tags = np.unique(l, return_inverse=True)
print(names) # ['alice' 'carter' 'michael']
print(tags) # [2, 2, 0, 1]
Hope this can help in the future.
