tags:

views:

140

answers:

4
Traceback (most recent call last):
  File "<pyshell#80>", line 1, in <module>
    do_work()
  File "C:\pythonwork\readthefile080410.py", line 14, in do_work
    populate_frequency5(e,data)
  File "C:\pythonwork\readthefile080410.py", line 157, in populate_frequency5
    data=medications_minimum3(data,[drug.upper()],1)
  File "C:\pythonwork\readthefile080410.py", line 120, in medications_minimum3
    counter[row[11]]+=1
TypeError: unhashable type: 'list'

i am getting this error on this line:

data=medications_minimum3(data,[drug.upper()],1)

(i have also tried drug.upper() without brackets

here is a preview of this function:

def medications_minimum3(c,drug_input,sample_cutoff): #return sample cut off for # medications/physician
  d=[]
  counter=collections.defaultdict(int)
  for row in c:
    counter[row[11]]+=1
  for row in c:
    if counter[row[11]]>=sample_cutoff:
      d.append(row) 
  write_file(d,'/pythonwork/medications_minimum3.csv')
  return d

does anyone know what am i doing wrong here?

i know that what must be wrong is the way i am calling this function, because i call this function from a different location and it works fine:

d=medications_minimum3(c,drug_input,50)

thank you very much for your help!

+3  A: 
counter[row[11]]+=1

You don't show what data is, but apparently when you loop through its rows, row[11] is turning out to be a list. Lists are mutable objects which means they cannot be used as dictionary keys. Trying to use row[11] as a key causes the defaultdict to complain that it is a mutable, i.e. unhashable, object.

The easiest fix is to change row[11] from a list to a tuple. Either by doing

counter[tuple(row[11])] += 1

or by fixing it in the caller before data is passed to medications_minimum3. A tuple simply an immutable list, so it behaves exactly like a list does except you cannot change it once it is created.

John Kugelman
thank very much for your help, the thing is as mentioned in the question, it is called from a different place and it works fine
I__
Right. If `c` contains tuples then the function works fine. If it contains lists then the function fails. It all depends on what the caller passes in.
John Kugelman
A: 
  File "C:\pythonwork\readthefile080410.py", line 120, in medications_minimum3
    counter[row[11]]+=1
TypeError: unhashable type: 'list'

row[11] is unhashable. It's a list. That is precisely (and only) what the error message means. You might not like it, but that is the error message.

Do this

counter[tuple(row[11])]+=1

Also, simplify.

d= [ row for row in c if counter[tuple(row[11])]>=sample_cutoff ]
S.Lott
the thing is as mentioned in the question, it is called from a different place and it works fine, counter is a collection
I__
"the thing is as mentioned in the question"? What? `counter` is a collection. Correct. `row[11]` is a list. That's what your error message says.
S.Lott
this is the wrong approach
I__
+1  A: 

As Jim Garrison said in the comment, no obvious reason why you'd make a one-element list out of drug.upper() (which implies drug is a string).

But that's not your error, as your function medications_minimum3() doesn't even use the second argument (something you should fix).

TypeError: unhashable type: 'list' usually means that you are trying to use a list as a hash argument (like for accessing a dictionary). I'd look for the error in counter[row[11]]+=1 -- are you sure that row[11] is of the right type? Sounds to me it might be a list.

chryss
+1  A: 

I don't think converting to a tuple is the right answer. You need go and look at where you are calling the function and make sure that c is a list of list of strings, or whatever you designed this function to work with

For example you might get this error if you passed [c] to the function instead of c

gnibbler