This code illustrates the use of partial application through the use of an operator section :
gt100 :: Integer -> Bool
gt100 x = x > 100
greaterThan100 :: [Integer] -> [Integer]
greaterThan100 xs = filter gt100 xs
greaterThan100_2 :: [Integer] -> [Integer]
greaterThan100_2 xs = filter (\x -> x > 100) xs
greaterThan100_3 :: [Integer] -> [Integer]
greaterThan100_3 xs = filter (>100) xs
The operator section (>100) partially applies the operator to one
of its two arguments. So the operator in this case is >
Why is this partial application as the > operator is being applied
to the entire Integer list ?
How is this different from lambda expression (\x -> x > 100) which
apparently is not partial application ?
Taken from http://www.seas.upenn.edu/~cis194/spring13/lectures/04-higher-order.html
Update :
Thanks to answers this appears clearer.
So here now is my understanding :
*Main> :t (>)
(>) :: Ord a => a -> a -> Bool
Is not applied at all as it accepts two parameters "a -> a" but they are not applied.
*Main> :t (>100)
(>100) :: (Ord a, Num a) => a -> Bool
Is partially applied as a function of type "a -> Bool" is created
*Main> :t (3>100)
(3>100) :: Bool
Is evaluated to type Bool which is the return type of operator (>) as illustrated by :t (>)
Why is this partial application as the > operator is being applied to the entire Integer list ?
It is not applied to the entire list, as a whole, but to the individual elements, one by one. When the > is partially applied over 100, it creates a new function.
Prelude> :type (> 100)
(> 100) :: (Num a, Ord a) => a -> Bool
Now, the function accepts an argument and returns Bool. This function is applied to all the elements in the list by the filter function.
How is this different from lambda expression (\x -> x > 100) which apparently is not partial application ?
> is a function which needs two operands to operate. You are already passing one of them, as 100. To completely execute the function, you need one more argument. So, the function > is partially applied with 100.
In the second case, you are creating a new function which needs only one argument to execute. Once you pass that argument, your function will be executed. So, the lambda function is not applied partially.
How is this [(<) 100] different from lambda expression (\x -> (<) 100 x) which apparently is not partial application? [1]
I think it is important to draw a distinction between semantic properties and syntactic properties. The two terms you suggested are semantically identical -- that is, there is no Haskell function you can write that reduces to one thing when applied to (<) 100 and reduces to a different thing when applied to (\x -> (<) 100 x).
However, partial application is a syntactic property -- that is, it is a property not of the behavior of a term but of the particular implementation chosen. (In analogous terms to my definition of a semantic property, a syntactic property is about whether you could write a Haskell function that reduces to one thing when applied to the string "(<) 100" and reduces to a different thing when applied to the string "(\x -> (<) 100 x)", which is eminently possible.) If I had to define it, I would define it this way: partial application is an application term with an arrow type. [2] Both of your terms are well-typed and have arrow types. But (<) 100 is an application term, whereas \x -> (<) 100 x is a lambda term (that has an application in its body).
As for operator sections like (100<) and (>100), it is not immediately obvious what to do with these. One choice is to simply decree that all operator sections are partial applications (or to simply decree that no operator sections are partial applications, of course). Another is to treat them as shorthand for (<) 100 and flip (>) 100, respectively, in which case we would still consider them partial applications (since they are application terms and have arrow types). [3] Yet a third is to treat them as shorthand for \x -> (<) 100 x and \x -> (>) x 100, respectively, in which case one might claim they are not partial applications.
But the distinction is, in my opinion, not too important: generally, semantic properties are more interesting and useful than merely syntactic ones.
[1] To avoid muddying the waters, I have used (<) 100 instead of (>100). We will discuss this distinction shortly.
[2] There is some question about what to do if the type is polymorphic. Let's punt on that question for now.
[3] This differs from simply decreeing that all operator sections are partial applications; consider some one-argument operator (!) and its section (100!), which we would then be treating as shorthand for (!) 100. This is an application term, but does not have an arrow in its type.
Related
I just want to know if my following understanding about currying functions is correct.
I want to filter all elements from a list which are > 4. I can achieve this like so:
filter (>4) [1..10]
(>) is defined as Ord a => a -> a -> Bool, thats why it cant be bassed to filter.
(>4) is defined as (Ord a, Num a) => a -> Bool. The function (>) is now curried and still expects one parameter.
Because of 2., (>4) can be passed to filter.
Every List-Element, which is passed to filter, will be passed to (>4) and and filter will validate the predicate and return the result.
Is this correct?
The reasoning is more or less correct. (>) is a function with signature:
(>) :: Ord a => a -> (a -> Bool)
so it is a function that (like any other function in Haskell) takes one parameter, and here returns a function a -> Bool.
The problem is thus that if we would use filter (>) [1,4,2,5], then we would make a call (>) 1, and this would thus return a function a -> Bool, but a filter can not work with that, it requires a function that takes an element from the list, and returns a Bool, not a function that returns a function that maps to a Bool.
We thus can for example use:
filter ((<) 4) [1,4,2,5]
Here we thus perform partial application [Haskell-wiki] of the function. This thus means that we generate a function (<) 4 :: (Num a, Ord a) => a -> Bool. We thus can filter with that function, it will thus return [5].
We can also make a function with a lambda expression to "swap" the order of the parameters:
filter (\x -> (>) x 4) [1,4,2,5]
Here for a value x, we will thus call ((>) x) 4, and this will thus return a Bool.
Since it happens often that one of the two sides of an infix operator is provided a value, Haskell has sectioning syntax for an infix operator [Haskell-wiki]:
(2^) (left section) is equivalent to (^) 2, or more verbosely \x -> 2 ^ x;
(^2) (right section) is equivalent to flip (^) 2, or more verbosely \x -> x ^ 2
So we can rewrite the last expression to:
filter (> 4) [1,4,2,5]
In working through a solution to the 8 Queens problem, a person used the following line of code:
sameDiag try qs = any (\(colDist,q) -> abs (try - q) == colDist) $ zip [1..] qs
try is an an item; qs is a list of the same items.
Can someone explain how colDist and q in the lambda function get bound to anything?
How did try and q used in the body of lambda function find their way into the same scope?
To the degree this is a Haskell idiom, what problem does this design approach help solve?
The function any is a higher-order function that takes 2 arguments:
the 1st argument is of type a -> Bool, i.e. a function from a to Bool
the 2nd argument is of type [a], i.e. a list of items of type a;
i.e. the 1st argument is a function that takes any element from the list passed as the 2nd argument, and returns a Bool based on that element. (well it can take any values of type a, not just the ones in that list, but it's quite obviously certain that any won't be invoking it with some arbitrary values of a but the ones from the list.)
You can then simplify thinking about the original snippet by doing a slight refactoring:
sameDiag :: Int -> [Int] -> Bool
sameDiag try qs = any f xs
where
xs = zip [1..] qs
f = (\(colDist, q) -> abs (try - q) == colDist)
which can be transformed into
sameDiag :: Int -> [Int] -> Bool
sameDiag try qs = any f xs
where
xs = zip [1..] qs
f (colDist, q) = abs (try - q) == colDist)
which in turn can be transformed into
sameDiag :: Int -> [Int] -> Bool
sameDiag try qs = any f xs
where
xs = zip [1..] qs
f pair = abs (try - q) == colDist) where (colDist, q) = pair
(Note that sameDiag could also have a more general type Integral a => a -> [a] -> Bool rather than the current monomorphic one)
— so how does the pair in f pair = ... get bound to a value? well, simple: it's just a function; whoever calls it must pass along a value for the pair argument. — when calling any with the first argument set to f, it's the invocation of the function any who's doing the calling of f, with individual elements of the list xs passed in as values of the argument pair.
and, since the contents of xs is a list of pairs, it's OK to pass an individual pair from this list to f as f expects it to be just that.
EDIT: a further explanation of any to address the asker's comment:
Is this a fair synthesis? This approach to designing a higher-order function allows the invoking code to change how f behaves AND invoke the higher-order function with a list that requires additional processing prior to being used to invoke f for every element in the list. Encapsulating the list processing (in this case with zip) seems the right thing to do, but is the intent of this additional processing really clear in the original one-liner above?
There's really no additional processing done by any prior to invoking f. There is just very minimalistic bookkeeping in addition to simply iterating through the passed in list xs: invoking f on the elements during the iteration, and immediately breaking the iteration and returning True the first time f returns True for any list element.
Most of the behavior of any is "implicit" though in that it's taken care of by Haskell's lazy evaluation, basic language semantics as well as existing functions, which any is composed of (well at least my version of it below, any' — I haven't taken a look at the built-in Prelude version of any yet but I'm sure it's not much different; just probably more heavily optimised).
In fact, any is simple it's almost trivial to re-implement it with a one liner on a GHCi prompt:
Prelude> let any' f xs = or (map f xs)
let's see now what GHC computes as its type:
Prelude> :t any'
any' :: (a -> Bool) -> [a] -> Bool
— same as the built-in any. So let's give it some trial runs:
Prelude> any' odd [1, 2, 3] -- any odd values in the list?
True
Prelude> any' even [1, 3] -- any even ones?
False
Prelude> let adult = (>=18)
Prelude> any' adult [17, 17, 16, 15, 17, 18]
— see how you can sometimes write code that almost looks like English with higher-order functions?
zip :: [a] -> [b] -> [(a,b)] takes two lists and joins them into pairs, dropping any remaining at the end.
any :: (a -> Bool) -> [a] -> Bool takes a function and a list of as and then returns True if any of the values returned true or not.
So colDist and q are the first and second elements of the pairs in the list made by zip [1..] qs, and they are bound when they are applied to the pair by any.
q is only bound within the body of the lambda function - this is the same as with lambda calculus. Since try was bound before in the function definition, it is still available in this inner scope. If you think of lambda calculus, the term \x.\y.x+y makes sense, despite the x and the y being bound at different times.
As for the design approach, this approach is much cleaner than trying to iterate or recurse through the list manually. It seems quite clear in its intentions to me (with respect to the larger codebase it comes from).
I've taken up learning Haskell again, after a short hiatus and I am currently trying to get a better understanding of how recursion and lambda expressions work in Haskell.
In this: YouTube video, there is a function example that puzzles me far more than it probably should, in terms of how it actually works:
firstThat :: (a -> Bool) -> a -> [a] -> a
firstThat f = foldr (\x acc -> if f x then x else acc)
For the sake of clarity and since it wasn't immediately obvious to me, I'll give an example of applying this function to some arguments:
firstThat (>10) 2000 [10,20,30,40] --returns 20, but would return 2000, if none of the values in the list were greater than 10
Please correct me, if my assumptions are wrong.
It seems firstThat takes three arguments:
a function that takes one arguments and returns a Boolean value. Since the > operator is actually an infix function, the first argument in the example above seems the result of a partial application to the > function – is this correct?
an unspecified value of the same type expected as the missing argument to the function provided as the first argument
a list of values of the aforementioned type
But the actual function firstThat seems to be defined differently from its type declaration, with just one argument. Since foldr normally takes three arguments I gathered there is some kind of partial application happening. The lambda expression provided as an argument to foldr seem to be missing its arguments too.
So, how exactly does this function work? I apologize if I am being too dense or fail to see the forest for the trees, but I just cannot wrap my head around it, which is frustrating.
Any helpful explanation or example(s) would be greatly appreciated.
Thanks!
But the actual function firstThat seems to be defined differently from its type declaration, with just one argument. Since foldr normally takes three arguments I gathered there is some kind of partial application happening.
You are right. However, there is a nicer way of putting it than talking about "missing arguments" -- one that doesn't lead you into asking where they have gone. Here are two ways in which the arguments are not missing.
Firstly, consider this function:
add :: Num a => a -> a -> a
add x y = x + y
As you may know, we can also define it like this:
add :: Num a => a -> a -> a
add = (+)
That works because Haskell functions are values like any other. We can simply define a value, add, as being equal to another value, (+), which just happens to be a function. There is no special syntax required to declare a function. The upshot is that writing arguments explicitly is (almost) never necessary; the main reason why we do so because it often makes code more readable (for instance, I could define firstThat without writing the f parameter explicitly, but I won't do so because the result is rather hideous).
Secondly, whenever you see a function type with three arguments...
firstThat :: (a -> Bool) -> a -> [a] -> a
... you can also read it like this...
firstThat :: (a -> Bool) -> (a -> [a] -> a)
... that is, a function of one argument that produces a function of two arguments. That works for all functions of more than one argument. The key takeaway is that, at heart, all Haskell functions take just one argument. That is why partial application works. So on seeing...
firstThat :: (a -> Bool) -> a -> [a] -> a
firstThat f = foldr (\x acc -> if f x then x else acc)
... you can accurately say that you have written explicitly all parameters that firstThat takes -- that is, only one :)
The lambda expression provided as an argument to foldr seem to be missing its arguments too.
Not really. foldr (when restricted to lists) is...
foldr :: (a -> b -> b) -> b -> [a] -> b
... and so the function passed to it takes two arguments (feel free to add air quotes around "two", given the discussion above). The lambda was written as...
\x acc -> if f x then x else acc
... with two explicit arguments, x and acc.
a function that takes one arguments and returns a Boolean value. Since the > operator is actually an infix function, the first argument in the example above seems the result of a partial application to the > function – is this correct?
yes: (>10) is short for \x -> x > 10, just as (10>) would be short for \x -> 10 > x.
an unspecified value of the same type expected as the missing argument to the function provided as the first argument
first of all, it's not a missing argument: by omitting an argument, you obtain a function value. however, the type of the 2nd argument does indeed match the argument of the function >10, just as it matches the type of the elements of the list [10,20,30,40] (which is better reasoning).
a list of values of the aforementioned type
yes.
But the actual function firstThat seems to be defined differently from its type declaration, with just one argument. Since foldr normally takes three arguments I gathered there is some kind of partial application happening. The lambda expression provided as an argument to foldr seem to be missing its arguments too.
that's because given e.g. foo x y z = x * y * z, these 2 lines are equivalent:
bar x = foo x
bar x y z = foo x y z
— that's because of a concept called currying. Currying is also the reason why function type signatures are not (a, b) -> c but instead a -> b -> c, which in turn is equivalent to a -> (b -> c) because of the right associativity of the -> type operator.
Therefore, these two lines are equivalent:
firstThat f = foldr (\x acc -> if f x then x else acc)
firstThat f x y = foldr (\x acc -> if f x then x else acc) x y
Note: that you can also use Data.List.find combined with Data.Maybe.fromMaybe:
λ> fromMaybe 2000 $ find (>10) [10, 20, 30]
20
λ> fromMaybe 2000 $ find (>10) [1, 2, 3]
2000
See also:
https://en.wikipedia.org/wiki/Currying.
https://www.fpcomplete.com/user/EFulmer/currying-and-partial-application
http://learnyouahaskell.com/higher-order-functions
In Haskell, some lists are cyclic:
ones = 1 : ones
Others are not:
nums = [1..]
And then there are things like this:
more_ones = f 1 where f x = x : f x
This denotes the same value as ones, and certainly that value is a repeating sequence. But whether it's represented in memory as a cyclic data structure is doubtful. (An implementation could do so, but this answer explains that "it's unlikely that this will happen in practice".)
Suppose we take a Haskell implementation and hack into it a built-in function isCycle :: [a] -> Bool that examines the structure of the in-memory representation of the argument. It returns True if the list is physically cyclic and False if the argument is of finite length. Otherwise, it will fail to terminate. (I imagine "hacking it in" because it's impossible to write that function in Haskell.)
Would the existence of this function break any interesting properties of the language?
Would the existence of this function break any interesting properties of the language?
Yes it would. It would break referential transparency (see also the Wikipedia article). A Haskell expression can be always replaced by its value. In other words, it depends only on the passed arguments and nothing else. If we had
isCycle :: [a] -> Bool
as you propose, expressions using it would not satisfy this property any more. They could depend on the internal memory representation of values. In consequence, other laws would be violated. For example the identity law for Functor
fmap id === id
would not hold any more: You'd be able to distinguish between ones and fmap id ones, as the latter would be acyclic. And compiler optimizations such as applying the above law would not longer preserve program properties.
However another question would be having function
isCycleIO :: [a] -> IO Bool
as IO actions are allowed to examine and change anything.
A pure solution could be to have a data type that internally distinguishes the two:
import qualified Data.Foldable as F
data SmartList a = Cyclic [a] | Acyclic [a]
instance Functor SmartList where
fmap f (Cyclic xs) = Cyclic (map f xs)
fmap f (Acyclic xs) = Acyclic (map f xs)
instance F.Foldable SmartList where
foldr f z (Acyclic xs) = F.foldr f z xs
foldr f _ (Cyclic xs) = let r = F.foldr f r xs in r
Of course it wouldn't be able to recognize if a generic list is cyclic or not, but for many operations it'd be possible to preserve the knowledge of having Cyclic values.
In the general case, no you can't identify a cyclic list. However if the list is being generated by an unfold operation then you can. Data.List contains this:
unfoldr :: (b -> Maybe (a, b)) -> b -> [a]
The first argument is a function that takes a "state" argument of type "b" and may return an element of the list and a new state. The second argument is the initial state. "Nothing" means the list ends.
If the state ever recurs then the list will repeat from the point of the last state. So if we instead use a different unfold function that returns a list of (a, b) pairs we can inspect the state corresponding to each element. If the same state is seen twice then the list is cyclic. Of course this assumes that the state is an instance of Eq or something.
I have to dessignate types of 2 functions(without using compiler :t) i just dont know how soudl i read these functions to make correct steps.
f x = map -1 x
f x = map (-1) x
Well i'm a bit confuse how it will be parsed
Function application, or "the empty space operator" has higher precedence than any operator symbol, so the first line parses as f x = map - (1 x), which will most likely1 be a type error.
The other example is parenthesized the way it looks, but note that (-1) desugars as negate 1. This is an exception from the normal rule, where operator sections like (+1) desugar as (\x -> x + 1), so this will also likely1 be a type error since map expects a function, not a number, as its first argument.
1 I say likely because it is technically possible to provide Num instances for functions which may allow this to type check.
For questions like this, the definitive answer is to check the Haskell Report. The relevant syntax hasn't changed from Haskell 98.
In particular, check the section on "Expressions". That should explain how expressions are parsed, operator precedence, and the like.
These functions do not have types, because they do not type check (you will get ridiculous type class constraints). To figure out why, you need to know that (-1) has type Num n => n, and you need to read up on how a - is interpreted with or without parens before it.
The following function is the "correct" version of your function:
f x = map (subtract 1) x
You should be able to figure out the type of this function, if I say that:
subtract 1 :: Num n => n -> n
map :: (a -> b) -> [a] -> [b]
well i did it by my self :P
(map) - (1 x)
(-)::Num a => a->a->->a
1::Num b=> b
x::e
map::(c->d)->[c]->[d]
map::a
a\(c->d)->[c]->[d]
(1 x)::a
1::e->a
f::(Num ((c->d)->[c]->[d]),Num (e->(c->d)->[c]->[d])) => e->(c->d)->[c]->[d]