Question

I have searched on strings, lists, append etc. but can’t seem to handle this.

I have created some files from android based on selections done in an app.
The output looks like this and is in a text file:

House 1,bwfront3,colorfront2,bwtilt3,colortilt3
House 2,bwfront6,colorfront6,bwtilt6,colortilt6
House 3,bwfront5,colorfront5,bwtilt5,colortilt5
House 4,bwfront4,colorfront4,bwtilt4,colortilt4
House 5,bwfront2,colorfront2,bwtilt2,colortilt2

the reason for the naming:

• I have 5 houses, where the user firstly selects from 9 ‘bwfront..’ pictures. and then between color images and so on.

• THe exercise is to map different pictures to the ‘house’.

• I now wish to load the text file(s) and count how many of each of the different ‘bwfront’ have been selected etc. To clarify, the user selects four times per ‘house’.

• This will continue with all houses + types of pictures,but if any of you can get me started, I should be able to apply the solution to all my 23 files.

Does it make sense?

Answer 1

# First, create a dictionary for each column, that maps each
# value (eg colorfront2) to a list of house names.
results = [{}, {}, {}, {}]
for filename in os.listdir('/wherever'):
    s = open(os.path.join('/wherever', filename), 'rb').read()
    for line in s.split('\n'):
        if not line.strip(): continue # skip blank lines
        house, values = line.split(',', 1)
        values = values.split(',')
        assert len(values) == len(results) # sanity check
        for value, result in zip(values, results):
            if value not in result:
                result[value] = []
            result[value].append(house)


# Then, do something with it -- e.g., show them in order
for i, result in enumerate(results):
    print 'COLUMN %d' % i
    def sortkey(pair): return len(pair[1]) # number of houses
    for (val, houses) in sorted(result.items(), key=sortkey, reverse=True):
        print '   %20s occurs for %d houses' % (val, len(houses))