CODEWITHSUNDEEP
haskell
Magnus Kronqvist
Magnus Kronqvist

Reputation: 1589

Making my own monadic type based on list

I want to use Haskell to solve a financial combinatorial problem, the list monad seems to be a good fit for this.

Now, my problem with the list monad is its inability to give names to the values involved. I will try to exemplify:

loan = [1000*x | x <- [1..3]]
interest_rate = [0.005*x | x <- [4..10]]

calc = do                                                                                    
  l <- loan                                                                                  
  i <- interest_rate                                                                         
  return (l*i)

Running calc above gives me a list of numbers ([20.0,25.0,30.0,35.0,40.0, ... ]), but I can't tell what the loan and interest rate is used for each calculation.

I get lost here, my intuition tells me to create my own monadic type of, say HelpfulNumber :: (String,[Double]) and somehow say that:

>>= and return should be >>= . snd and return . snd

Am I on the right course here, or is there a better way? I am feeling a bit lost to be honest.

Upvotes: 3

Views: 201

Answers (3)

AndrewC
AndrewC

Reputation: 32455

You could use a record type to make your output clearer:

data Loan = Loan {final :: Double, 
                  rate  :: Double, 
                  loan  :: Integer, 
                  years :: Int}
   deriving Show

printloans :: [Loan] -> IO()
printloans = mapM_ print

Use printloans loans or printloans loans' at the ghci prompt.

Edit: I forgot to include the definition of dp. It's for rounding to a given number of decimal places:

dp :: Int -> Double -> Double
n `dp` a = (/ 10.0^n).fromInteger.round.(* 10.0^n) $ a

Here's a way using a list directly:

loans = [Loan {final = (2 `dp`) $ fromInteger amt*(1+ir)^yrs, 
               rate  = ir,
               loan  = amt,
               years = yrs}  
        | ir <- [0.005*x | x <- [4..10]],
          amt <- [1000*x | x <- [1..3]],
          yrs <- [1..4]
        ]

But if you like the monadic style, you can use:

loans' = do
  ir <- [0.005*x | x <- [4..10]]
  amt <- [1000*x | x <- [1..3]]
  yrs <- [1..4]
  return Loan {final = (2 `dp`) $ fromInteger amt*(1+ir)^yrs, 
               rate  = ir,
               loan  = amt,
               years = yrs}

which benefits from fewer commas, and it's easier to change the order of the <- lines to change the order of the answers. You can add extras to your Loan record and calculate with them. You get output like this:

*Main> printloans loans'
Loan {final = 1020.0, rate = 2.0e-2, loan = 1000, years = 1}
Loan {final = 1040.4, rate = 2.0e-2, loan = 1000, years = 2}
Loan {final = 1061.21, rate = 2.0e-2, loan = 1000, years = 3}
Loan {final = 1082.43, rate = 2.0e-2, loan = 1000, years = 4}
Loan {final = 2040.0, rate = 2.0e-2, loan = 2000, years = 1}
Loan {final = 2080.8, rate = 2.0e-2, loan = 2000, years = 2}
...
...

EDIT:

You told me elsewhere you'd like output like ir_5% yrs_3 amt_4000 tot_4360.5. It's uglier, but here's a way of doing that sort of thing:

loans'' = do
  ir <- [0.005*x | x <- [4..10]]
  amt <- [1000*x | x <- [1..3]]
  yrs <- [1..4]
  let final = (2 `dp`) $ fromInteger amt*(1+ir)^yrs
  return $ "final_" ++ show final
        ++ ",  ir_" ++ show ((2 `dp`) $ ir*100.0)    -- rounded away a rounding error in 3.5% 
        ++ "%,  amt_" ++ show amt 
        ++ ",  yrs_" ++ show yrs

When you do mapM_ putStrLn loans'' you get output like

final_1020.0,  ir_2.0%,  amt_1000,  yrs_1
final_1040.4,  ir_2.0%,  amt_1000,  yrs_2
final_1061.21,  ir_2.0%,  amt_1000,  yrs_3
final_1082.43,  ir_2.0%,  amt_1000,  yrs_4
final_2040.0,  ir_2.0%,  amt_2000,  yrs_1
....

but I think the record type is much nicer - its output is easier to read and there's less messing about with strings.

Upvotes: 6

Dax Fohl
Dax Fohl

Reputation: 10781

Why don't you just make two helper functions?

getName (loan, rate) = "loan="++loan++"&rate="++rate

getAnswer (loan, rate) = loan*rate

Then use a list comprehension

loans = [1000*x | x <- [1..3]]
interest_rates = [0.005*x | x <- [4..10]]

input_tuples = [(l, i) | l<-loans, i<-interest_rates]]

answers = [(getName t, getAnswer t) | t<-input_tuples]]

No monads necessary, not even list monad.

Upvotes: 2

dflemstr
dflemstr

Reputation: 26167

I don't know exactly what you're looking for here, because how would you give names to interest rates?

But you can of course store the interest rate along with the final result:

calc = do                                                                                    
  l <- loan                                                                                  
  i <- interest_rate                                                                         
  return (i, l*i)
-- Yields: [(0.02, 20.0), (0.025, 25.0), ...]

Upvotes: 1

Related Questions