I want to write a tail recursive function to multiply all the values in a list by 2 in F#. I know there is a bunch of ways to do this but i want to know if this is even a viable method. This is purely for educational purposes. I realize that there is a built in function to do this for me.
let multiply m =
let rec innerfunct ax = function
| [] -> printfn "%A" m
| (car::cdr) -> (car <- car*2 innerfunct cdr);
innerfunct m;;
let mutable a = 1::3::4::[]
multiply a
I get two errors with this though i doubt they are the only problems.
This value is not mutable on my second matching condition
and
This expression is a function value, i.e. is missing arguments. Its type is 'a list -> unit. for when i call length a.
I am fairly new to F# and realize im probably not calling the function properly but i cant figure out why. This is mostly a learning experience for me so the explanation is more important than just fixing the code. The syntax is clearly off, but can i map *2 to a list just by doing the equivalent of
car = car*2 and then calling the inner function on the cdr of the list.
There are a number of issues that I can't easily explain without showing intermediate code, so I'll try to walk through a commented refactoring:
First, we'll go down the mutable path:
As F# lists are immutable and so are primitive ints, we need a way to mutate that thing inside the list:
let mutable a = [ref 1; ref 3; ref 4]
Getting rid of the superfluous ax and arranging the cases a bit, we can make use of these reference cells:
let multiply m =
let rec innerfunct = function
| [] -> printfn "%A" m
| car :: cdr ->
car := !car*2
innerfunct cdr
innerfunct m
We see, that multiply only calls its inner function, so we end up with the first solution:
let rec multiply m =
match m with
| [] -> printfn "%A" m
| car :: cdr ->
car := !car*2
multiply cdr
This is really only for it's own purpose. If you want mutability, use arrays and traditional for-loops.
Then, we go up the immutable path:
As we learnt in the mutable world, the first error is due to car not being mutable. It is just a primitive int out of an immutable list. Living in an immutable world means we can only create something new out of our input. What we want is to construct a new list, having car*2 as head and then the result of the recursive call to innerfunct. As usual, all branches of a function need to return some thing of the same type:
let multiply m =
let rec innerfunct = function
| [] ->
printfn "%A" m
[]
| car :: cdr ->
car*2 :: innerfunct cdr
innerfunct m
Knowing m is immutable, we can get rid of the printfn. If needed, we can put it outside of the function, anywhere we have access to the list. It will always print the same.
We finish by also making the reference to the list immutable and obtain a second (intermediate) solution:
let multiply m =
let rec innerfunct = function
| [] -> []
| car :: cdr -> car*2 :: innerfunct cdr
innerfunct m
let a = [1; 3; 4]
printfn "%A" a
let multiplied = multiply a
printfn "%A" multiplied
It might be nice to also multiply by different values (the function is called multiply after all and not double). Also, now that innerfunct is so small, we can make the names match the small scope (the smaller the scope, the shorter the names):
let multiply m xs =
let rec inner = function
| [] -> []
| x :: tail -> x*m :: inner tail
inner xs
Note that I put the factor first and the list last. This is similar to other List functions and allows to create pre-customized functions by using partial application:
let double = multiply 2
let doubled = double a
All that's left now is to make multiply tail-recursive:
let multiply m xs =
let rec inner acc = function
| [] -> acc
| x :: tail -> inner (x*m :: acc) tail
inner [] xs |> List.rev
So we end up having (for educational purposes) a hard-coded version of let multiply' m = List.map ((*) m)
F# is a 'single-pass' compiler, so you can expect any compilation error to have a cascading effect beneath the error. When you have a compilation error, focus on that single error. While you may have more errors in your code (you do), it may also be that subsequent errors are only consequences of the first error.
As the compiler says, car isn't mutable, so you can assign a value to it.
In Functional Programming, a map can easily be implemented as a recursive function:
// ('a -> 'b) -> 'a list -> 'b list
let rec map f = function
| [] -> []
| h::t -> f h :: map f t
This version, however, isn't tail-recursive, since it recursively calls map before it cons the head onto the tail.
You can normally refactor to a tail-recursive implementation by introducing an 'inner' implementation function that uses an accumulator for the result. Here's one way to do that:
// ('a -> 'b) -> 'a list -> 'b list
let map' f xs =
let rec mapImp f acc = function
| [] -> acc
| h::t -> mapImp f (acc # [f h]) t
mapImp f [] xs
Here, mapImp is the last operation to be invoked in the h::t case.
This implementation is a bit inefficient because it concatenates two lists (acc # [f h]) in each iteration. Depending on the size of the lists to map, it may be more efficient to cons the accumulator and then do a single reverse at the end:
// ('a -> 'b) -> 'a list -> 'b list
let map'' f xs =
let rec mapImp f acc = function
| [] -> acc
| h::t -> mapImp f (f h :: acc) t
mapImp f [] xs |> List.rev
In any case, however, the only reason to do all of this is for the exercise, because this function is already built-in.
In all cases, you can use map functions to multiply all elements in a list by two:
> let mdouble = List.map ((*) 2);;
val mdouble : (int list -> int list)
> mdouble [1..10];;
val it : int list = [2; 4; 6; 8; 10; 12; 14; 16; 18; 20]
Normally, though, I wouldn't even care to define such function explicitly. Instead, you use it inline:
> List.map ((*) 2) [1..10];;
val it : int list = [2; 4; 6; 8; 10; 12; 14; 16; 18; 20]
You can use all the above map function in the same way.
Symbols that you are creating in a match statement are not mutable, so when you are matching with (car::cdr) you cannot change their values.
Standard functional way would be to produce a new list with the computed values. For that you can write something like this:
let multiplyBy2 = List.map (fun x -> x * 2)
multiplyBy2 [1;2;3;4;5]
This is not tail recursive by itself (but List.map is).
If you really want to change values of the list, you could use an array instead. Then your function will not produce any new objects, just iterate through the array:
let multiplyArrayBy2 arr =
arr
|> Array.iteri (fun index value -> arr.[index] <- value * 2)
let someArray = [| 1; 2; 3; 4; 5 |]
multiplyArrayBy2 someArray
I have a computation expression builder that builds up a value as you go, and has many custom operations. However, it does not allow for standard F# language constructs, and I'm having a lot of trouble figuring out how to add this support.
To give a stand-alone example, here's a dead-simple and fairly pointless computation expression that builds F# lists:
type Items<'a> = Items of 'a list
type ListBuilder() =
member x.Yield(()) = Items []
[<CustomOperation("add")>]
member x.Add(Items current, item:'a) =
Items [ yield! current; yield item ]
[<CustomOperation("addMany")>]
member x.AddMany(Items current, items: seq<'a>) =
Items [ yield! current; yield! items ]
let listBuilder = ListBuilder()
let build (Items items) = items
I can use this to build lists just fine:
let stuff =
listBuilder {
add 1
add 5
add 7
addMany [ 1..10 ]
add 42
}
|> build
However, this is a compiler error:
listBuilder {
let x = 5 * 39
add x
}
// This expression was expected to have type unit, but
// here has type int.
And so is this:
listBuilder {
for x = 1 to 50 do
add x
}
// This control construct may only be used if the computation expression builder
// defines a For method.
I've read all the documentation and examples I can find, but there's something I'm just not getting. Every .Bind() or .For() method signature I try just leads to more and more confusing compiler errors. Most of the examples I can find either build up a value as you go along, or allow for regular F# language constructs, but I haven't been able to find one that does both.
If someone could point me in the right direction by showing me how to take this example and add support in the builder for let bindings and for loops (at minimum - using, while and try/catch would be great, but I can probably figure those out if someone gets me started) then I'll be able to gratefully apply the lesson to my actual problem.
The best place to look is the spec. For example,
b {
let x = e
op x
}
gets translated to
T(let x = e in op x, [], fun v -> v, true)
=> T(op x, {x}, fun v -> let x = e in v, true)
=> [| op x, let x = e in b.Yield(x) |]{x}
=> b.Op(let x = e in in b.Yield(x), x)
So this shows where things have gone wrong, though it doesn't present an obvious solution. Clearly, Yield needs to be generalized since it needs to take arbitrary tuples (based on how many variables are in scope). Perhaps more subtly, it also shows that x is not in scope in the call to add (see that unbound x as the second argument to b.Op?). To allow your custom operators to use bound variables, their arguments need to have the [<ProjectionParameter>] attribute (and take functions from arbitrary variables as arguments), and you'll also need to set MaintainsVariableSpace to true if you want bound variables to be available to later operators. This will change the final translation to:
b.Op(let x = e in b.Yield(x), fun x -> x)
Building up from this, it seems that there's no way to avoid passing the set of bound values along to and from each operation (though I'd love to be proven wrong) - this will require you to add a Run method to strip those values back off at the end. Putting it all together, you'll get a builder which looks like this:
type ListBuilder() =
member x.Yield(vars) = Items [],vars
[<CustomOperation("add",MaintainsVariableSpace=true)>]
member x.Add((Items current,vars), [<ProjectionParameter>]f) =
Items (current # [f vars]),vars
[<CustomOperation("addMany",MaintainsVariableSpace=true)>]
member x.AddMany((Items current, vars), [<ProjectionParameter>]f) =
Items (current # f vars),vars
member x.Run(l,_) = l
The most complete examples I've seen are in §6.3.10 of the spec, especially this one:
/// Computations that can cooperatively yield by returning a continuation
type Eventually<'T> =
| Done of 'T
| NotYetDone of (unit -> Eventually<'T>)
[<CompilationRepresentation(CompilationRepresentationFlags.ModuleSuffix)>]
module Eventually =
/// The bind for the computations. Stitch 'k' on to the end of the computation.
/// Note combinators like this are usually written in the reverse way,
/// for example,
/// e |> bind k
let rec bind k e =
match e with
| Done x -> NotYetDone (fun () -> k x)
| NotYetDone work -> NotYetDone (fun () -> bind k (work()))
/// The return for the computations.
let result x = Done x
type OkOrException<'T> =
| Ok of 'T
| Exception of System.Exception
/// The catch for the computations. Stitch try/with throughout
/// the computation and return the overall result as an OkOrException.
let rec catch e =
match e with
| Done x -> result (Ok x)
| NotYetDone work ->
NotYetDone (fun () ->
let res = try Ok(work()) with | e -> Exception e
match res with
| Ok cont -> catch cont // note, a tailcall
| Exception e -> result (Exception e))
/// The delay operator.
let delay f = NotYetDone (fun () -> f())
/// The stepping action for the computations.
let step c =
match c with
| Done _ -> c
| NotYetDone f -> f ()
// The rest of the operations are boilerplate.
/// The tryFinally operator.
/// This is boilerplate in terms of "result", "catch" and "bind".
let tryFinally e compensation =
catch (e)
|> bind (fun res -> compensation();
match res with
| Ok v -> result v
| Exception e -> raise e)
/// The tryWith operator.
/// This is boilerplate in terms of "result", "catch" and "bind".
let tryWith e handler =
catch e
|> bind (function Ok v -> result v | Exception e -> handler e)
/// The whileLoop operator.
/// This is boilerplate in terms of "result" and "bind".
let rec whileLoop gd body =
if gd() then body |> bind (fun v -> whileLoop gd body)
else result ()
/// The sequential composition operator
/// This is boilerplate in terms of "result" and "bind".
let combine e1 e2 =
e1 |> bind (fun () -> e2)
/// The using operator.
let using (resource: #System.IDisposable) f =
tryFinally (f resource) (fun () -> resource.Dispose())
/// The forLoop operator.
/// This is boilerplate in terms of "catch", "result" and "bind".
let forLoop (e:seq<_>) f =
let ie = e.GetEnumerator()
tryFinally (whileLoop (fun () -> ie.MoveNext())
(delay (fun () -> let v = ie.Current in f v)))
(fun () -> ie.Dispose())
// Give the mapping for F# computation expressions.
type EventuallyBuilder() =
member x.Bind(e,k) = Eventually.bind k e
member x.Return(v) = Eventually.result v
member x.ReturnFrom(v) = v
member x.Combine(e1,e2) = Eventually.combine e1 e2
member x.Delay(f) = Eventually.delay f
member x.Zero() = Eventually.result ()
member x.TryWith(e,handler) = Eventually.tryWith e handler
member x.TryFinally(e,compensation) = Eventually.tryFinally e compensation
member x.For(e:seq<_>,f) = Eventually.forLoop e f
member x.Using(resource,e) = Eventually.using resource e
The tutorial at "F# for fun and profit" is first class in this regard.
http://fsharpforfunandprofit.com/posts/computation-expressions-intro/
Following a similar struggle to Joel's (and not finding §6.3.10 of the spec that helpful) my issue with getting the For construct to generate a list came down to getting types to line up properly (no special attributes required). In particular I was slow to realise that For would build a list of lists, and therefore need flattening, despite the best efforts of the compiler to put me right. Examples that I found on the web were always wrappers around seq{}, using the yield keyword, repeated use of which invokes a call to Combine, which does the flattening. In case a concrete example helps, the following excerpt uses for to build a list of integers - my ultimate aim being to create lists of components for rendering in a GUI (with some additional laziness thrown in). Also In depth talk on CE here which elaborates on kvb's points above.
module scratch
type Dispatcher = unit -> unit
type viewElement = int
type lazyViews = Lazy<list<viewElement>>
type ViewElementsBuilder() =
member x.Return(views: lazyViews) : list<viewElement> = views.Value
member x.Yield(v: viewElement) : list<viewElement> = [v]
member x.ReturnFrom(viewElements: list<viewElement>) = viewElements
member x.Zero() = list<viewElement>.Empty
member x.Combine(listA:list<viewElement>, listB: list<viewElement>) = List.concat [listA; listB]
member x.Delay(f) = f()
member x.For(coll:seq<'a>, forBody: 'a -> list<viewElement>) : list<viewElement> =
// seq {for v in coll do yield! f v} |> List.ofSeq
Seq.map forBody coll |> Seq.collect id |> List.ofSeq
let ve = new ViewElementsBuilder()
let makeComponent(m: int, dispatch: Dispatcher) : viewElement = m
let makeComponents() : list<viewElement> = [77; 33]
let makeViewElements() : list<viewElement> =
let model = {| Scores = [33;23;22;43;] |> Seq.ofList; Trainer = "John" |}
let d:Dispatcher = fun() -> () // Does nothing here, but will be used to raise messages from UI
ve {
for score in model.Scores do
yield makeComponent (score, d)
yield makeComponent (score * 100 / 50 , d)
if model.Trainer = "John" then
return lazy
[ makeComponent (12, d)
makeComponent (13, d)
]
else
return lazy
[ makeComponent (14, d)
makeComponent (15, d)
]
yield makeComponent (33, d)
return! makeComponents()
}
As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, visit the help center for guidance.
Closed 10 years ago.
There are already two questions about F#/functional snippets.
However what I'm looking for here are useful snippets, little 'helper' functions that are reusable. Or obscure but nifty patterns that you can never quite remember.
Something like:
open System.IO
let rec visitor dir filter=
seq { yield! Directory.GetFiles(dir, filter)
for subdir in Directory.GetDirectories(dir) do
yield! visitor subdir filter}
I'd like to make this a kind of handy reference page. As such there will be no right answer, but hopefully lots of good ones.
EDIT Tomas Petricek has created a site specifically for F# snippets http://fssnip.net/.
Perl style regex matching
let (=~) input pattern =
System.Text.RegularExpressions.Regex.IsMatch(input, pattern)
It lets you match text using let test = "monkey" =~ "monk.+" notation.
Infix Operator
I got this from http://sandersn.com/blog//index.php/2009/10/22/infix-function-trick-for-f go to that page for more details.
If you know Haskell, you might find yourself missing infix sugar in F#:
// standard Haskell call has function first, then args just like F#. So obviously
// here there is a function that takes two strings: string -> string -> string
startsWith "kevin" "k"
//Haskell infix operator via backQuotes. Sometimes makes a function read better.
"kevin" `startsWith` "K"
While F# doesn't have a true 'infix' operator, the same thing can be accomplished almost as elegantly via a pipeline and a 'backpipeline' (who knew of such a thing??)
// F# 'infix' trick via pipelines
"kevin" |> startsWith <| "K"
Multi-Line Strings
This is pretty trivial, but it seems to be a feature of F# strings that is not widely known.
let sql = "select a,b,c \
from table \
where a = 1"
This produces:
val sql : string = "select a,b,c from table where a = 1"
When the F# compiler sees a back-slash followed by a carriage return inside a string literal, it will remove everything from the back-slash to the first non-space character on the next line. This allows you to have multi-line string literals that line up, without using a bunch of string concatenation.
Generic memoization, courtesy of the man himself
let memoize f =
let cache = System.Collections.Generic.Dictionary<_,_>(HashIdentity.Structural)
fun x ->
let ok, res = cache.TryGetValue(x)
if ok then res
else let res = f x
cache.[x] <- res
res
Using this, you could do a cached reader like so:
let cachedReader = memoize reader
Simple read-write to text files
These are trivial, but make file access pipeable:
open System.IO
let fileread f = File.ReadAllText(f)
let filewrite f s = File.WriteAllText(f, s)
let filereadlines f = File.ReadAllLines(f)
let filewritelines f ar = File.WriteAllLines(f, ar)
So
let replace f (r:string) (s:string) = s.Replace(f, r)
"C:\\Test.txt" |>
fileread |>
replace "teh" "the" |>
filewrite "C:\\Test.txt"
And combining that with the visitor quoted in the question:
let filereplace find repl path =
path |> fileread |> replace find repl |> filewrite path
let recurseReplace root filter find repl =
visitor root filter |> Seq.iter (filereplace find repl)
Update Slight improvement if you want to be able to read 'locked' files (e.g. csv files which are already open in Excel...):
let safereadall f =
use fs = new FileStream(f, FileMode.Open, FileAccess.Read, FileShare.ReadWrite)
use sr = new StreamReader(fs, System.Text.Encoding.Default)
sr.ReadToEnd()
let split sep (s:string) = System.Text.RegularExpressions.Regex.Split(s, sep)
let fileread f = safereadall f
let filereadlines f = f |> safereadall |> split System.Environment.NewLine
For performance intensive stuff where you need to check for null
let inline isNull o = System.Object.ReferenceEquals(o, null)
if isNull o then ... else ...
Is about 20x faster then
if o = null then ... else ...
Active Patterns, aka "Banana Splits", are a very handy construct that let one match against multiple regular expression patterns. This is much like AWK, but without the high performance of DFA's because the patterns are matched in sequence until one succeeds.
#light
open System
open System.Text.RegularExpressions
let (|Test|_|) pat s =
if (new Regex(pat)).IsMatch(s)
then Some()
else None
let (|Match|_|) pat s =
let opt = RegexOptions.None
let re = new Regex(pat,opt)
let m = re.Match(s)
if m.Success
then Some(m.Groups)
else None
Some examples of use:
let HasIndefiniteArticle = function
| Test "(?: |^)(a|an)(?: |$)" _ -> true
| _ -> false
type Ast =
| IntVal of string * int
| StringVal of string * string
| LineNo of int
| Goto of int
let Parse = function
| Match "^LET\s+([A-Z])\s*=\s*(\d+)$" g ->
IntVal( g.[1].Value, Int32.Parse(g.[2].Value) )
| Match "^LET\s+([A-Z]\$)\s*=\s*(.*)$" g ->
StringVal( g.[1].Value, g.[2].Value )
| Match "^(\d+)\s*:$" g ->
LineNo( Int32.Parse(g.[1].Value) )
| Match "^GOTO \s*(\d+)$" g ->
Goto( Int32.Parse(g.[1].Value) )
| s -> failwithf "Unexpected statement: %s" s
Maybe monad
type maybeBuilder() =
member this.Bind(v, f) =
match v with
| None -> None
| Some(x) -> f x
member this.Delay(f) = f()
member this.Return(v) = Some v
let maybe = maybeBuilder()
Here's a brief intro to monads for the uninitiated.
Option-coalescing operators
I wanted a version of the defaultArg function that had a syntax closer to the C# null-coalescing operator, ??. This lets me get the value from an Option while providing a default value, using a very concise syntax.
/// Option-coalescing operator - this is like the C# ?? operator, but works with
/// the Option type.
/// Warning: Unlike the C# ?? operator, the second parameter will always be
/// evaluated.
/// Example: let foo = someOption |? default
let inline (|?) value defaultValue =
defaultArg value defaultValue
/// Option-coalescing operator with delayed evaluation. The other version of
/// this operator always evaluates the default value expression. If you only
/// want to create the default value when needed, use this operator and pass
/// in a function that creates the default.
/// Example: let foo = someOption |?! (fun () -> new Default())
let inline (|?!) value f =
match value with Some x -> x | None -> f()
'Unitize' a function which doesn't handle units
Using the FloatWithMeasure function http://msdn.microsoft.com/en-us/library/ee806527(VS.100).aspx.
let unitize (f:float -> float) (v:float<'u>) =
LanguagePrimitives.FloatWithMeasure<'u> (f (float v))
Example:
[<Measure>] type m
[<Measure>] type kg
let unitize (f:float -> float) (v:float<'u>) =
LanguagePrimitives.FloatWithMeasure<'u> (f (float v))
//this function doesn't take units
let badinc a = a + 1.
//this one does!
let goodinc v = unitize badinc v
goodinc 3.<m>
goodinc 3.<kg>
OLD version:
let unitize (f:float -> float) (v:float<'u>) =
let unit = box 1. :?> float<'u>
unit * (f (v/unit))
Kudos to kvb
Scale/Ratio function builder
Again, trivial, but handy.
//returns a function which will convert from a1-a2 range to b1-b2 range
let scale (a1:float<'u>, a2:float<'u>) (b1:float<'v>,b2:float<'v>) =
let m = (b2 - b1)/(a2 - a1) //gradient of line (evaluated once only..)
(fun a -> b1 + m * (a - a1))
Example:
[<Measure>] type m
[<Measure>] type px
let screenSize = (0.<px>, 300.<px>)
let displayRange = (100.<m>, 200.<m>)
let scaleToScreen = scale displayRange screenSize
scaleToScreen 120.<m> //-> 60.<px>
Transposing a list (seen on Jomo Fisher's blog)
///Given list of 'rows', returns list of 'columns'
let rec transpose lst =
match lst with
| (_::_)::_ -> List.map List.head lst :: transpose (List.map List.tail lst)
| _ -> []
transpose [[1;2;3];[4;5;6];[7;8;9]] // returns [[1;4;7];[2;5;8];[3;6;9]]
And here is a tail-recursive version which (from my sketchy profiling) is mildly slower, but has the advantage of not throwing a stack overflow when the inner lists are longer than 10000 elements (on my machine):
let transposeTR lst =
let rec inner acc lst =
match lst with
| (_::_)::_ -> inner (List.map List.head lst :: acc) (List.map List.tail lst)
| _ -> List.rev acc
inner [] lst
If I was clever, I'd try and parallelise it with async...
F# Map <-> C# Dictionary
(I know, I know, System.Collections.Generic.Dictionary isn't really a 'C#' dictionary)
C# to F#
(dic :> seq<_>) //cast to seq of KeyValuePair
|> Seq.map (|KeyValue|) //convert KeyValuePairs to tuples
|> Map.ofSeq //convert to Map
(From Brian, here, with improvement proposed by Mauricio in comment below. (|KeyValue|) is an active pattern for matching KeyValuePair - from FSharp.Core - equivalent to (fun kvp -> kvp.Key, kvp.Value))
Interesting alternative
To get all of the immutable goodness, but with the O(1) lookup speed of Dictionary, you can use the dict operator, which returns an immutable IDictionary (see this question).
I currently can't see a way to directly convert a Dictionary using this method, other than
(dic :> seq<_>) //cast to seq of KeyValuePair
|> (fun kvp -> kvp.Key, kvp.Value) //convert KeyValuePairs to tuples
|> dict //convert to immutable IDictionary
F# to C#
let dic = Dictionary()
map |> Map.iter (fun k t -> dic.Add(k, t))
dic
What is weird here is that FSI will report the type as (for example):
val it : Dictionary<string,int> = dict [("a",1);("b",2)]
but if you feed dict [("a",1);("b",2)] back in, FSI reports
IDictionary<string,int> = seq[[a,1] {Key = "a"; Value = 1; } ...
Tree-sort / Flatten a tree into a list
I have the following binary tree:
___ 77 _
/ \
______ 47 __ 99
/ \
21 _ 54
\ / \
43 53 74
/
39
/
32
Which is represented as follows:
type 'a tree =
| Node of 'a tree * 'a * 'a tree
| Nil
let myTree =
Node
(Node
(Node (Nil,21,Node (Node (Node (Nil,32,Nil),39,Nil),43,Nil)),47,
Node (Node (Nil,53,Nil),54,Node (Nil,74,Nil))),77,Node (Nil,99,Nil))
A straightforward method to flatten the tree is:
let rec flatten = function
| Nil -> []
| Node(l, a, r) -> flatten l # a::flatten r
This isn't tail-recursive, and I believe the # operator causes it to be O(n log n) or O(n^2) with unbalanced binary trees. With a little tweaking, I came up with this tail-recursive O(n) version:
let flatten2 t =
let rec loop acc c = function
| Nil -> c acc
| Node(l, a, r) ->
loop acc (fun acc' -> loop (a::acc') c l) r
loop [] (fun x -> x) t
Here's the output in fsi:
> flatten2 myTree;;
val it : int list = [21; 32; 39; 43; 47; 53; 54; 74; 77; 99]
LINQ-to-XML helpers
namespace System.Xml.Linq
// hide warning about op_Explicit
#nowarn "77"
[<AutoOpen>]
module XmlUtils =
/// Converts a string to an XName.
let xn = XName.op_Implicit
/// Converts a string to an XNamespace.
let xmlns = XNamespace.op_Implicit
/// Gets the string value of any XObject subclass that has a Value property.
let inline xstr (x : ^a when ^a :> XObject) =
(^a : (member get_Value : unit -> string) x)
/// Gets a strongly-typed value from any XObject subclass, provided that
/// an explicit conversion to the output type has been defined.
/// (Many explicit conversions are defined on XElement and XAttribute)
/// Example: let value:int = xval foo
let inline xval (x : ^a when ^a :> XObject) : ^b =
((^a or ^b) : (static member op_Explicit : ^a -> ^b) x)
/// Dynamic lookup operator for getting an attribute value from an XElement.
/// Returns a string option, set to None if the attribute was not present.
/// Example: let value = foo?href
/// Example with default: let value = defaultArg foo?Name "<Unknown>"
let (?) (el:XElement) (name:string) =
match el.Attribute(xn name) with
| null -> None
| att -> Some(att.Value)
/// Dynamic operator for setting an attribute on an XElement.
/// Example: foo?href <- "http://www.foo.com/"
let (?<-) (el:XElement) (name:string) (value:obj) =
el.SetAttributeValue(xn name, value)
OK, this has nothing to do with snippets, but I keep forgetting this:
If you are in the interactive window, you hit F7 to jump back to the code window (without deselecting the code which you just ran...)
Going from code window to F# window (and also to open the F# window) is Ctrl Alt F
(unless CodeRush has stolen your bindings...)
Weighted sum of arrays
Calculating a weighted [n-array] sum of a [k-array of n-arrays] of numbers, based on a [k-array] of weights
(Copied from this question, and kvb's answer)
Given these arrays
let weights = [|0.6;0.3;0.1|]
let arrs = [| [|0.0453;0.065345;0.07566;1.562;356.6|] ;
[|0.0873;0.075565;0.07666;1.562222;3.66|] ;
[|0.06753;0.075675;0.04566;1.452;3.4556|] |]
We want a weighted sum (by column), given that both dimensions of the arrays can be variable.
Array.map2 (fun w -> Array.map ((*) w)) weights arrs
|> Array.reduce (Array.map2 (+))
First line: Partial application of the first Array.map2 function to weights yields a new function (Array.map ((*) weight) which is applied (for each weight) to each array in arr.
Second line: Array.reduce is like fold, except it starts on the second value and uses the first as the initial 'state'. In this case each value is a 'line' of our array of arrays. So applying an Array.map2 (+) on the first two lines means that we sum the first two arrays, which leaves us with a new array, which we then (Array.reduce) sum again onto the next (in this case last) array.
Result:
[|0.060123; 0.069444; 0.07296; 1.5510666; 215.40356|]
Performance testing
(Found here and updated for latest release of F#)
open System
open System.Diagnostics
module PerformanceTesting =
let Time func =
let stopwatch = new Stopwatch()
stopwatch.Start()
func()
stopwatch.Stop()
stopwatch.Elapsed.TotalMilliseconds
let GetAverageTime timesToRun func =
Seq.initInfinite (fun _ -> (Time func))
|> Seq.take timesToRun
|> Seq.average
let TimeOperation timesToRun =
GC.Collect()
GetAverageTime timesToRun
let TimeOperations funcsWithName =
let randomizer = new Random(int DateTime.Now.Ticks)
funcsWithName
|> Seq.sortBy (fun _ -> randomizer.Next())
|> Seq.map (fun (name, func) -> name, (TimeOperation 100000 func))
let TimeOperationsAFewTimes funcsWithName =
Seq.initInfinite (fun _ -> (TimeOperations funcsWithName))
|> Seq.take 50
|> Seq.concat
|> Seq.groupBy fst
|> Seq.map (fun (name, individualResults) -> name, (individualResults |> Seq.map snd |> Seq.average))
DataSetExtensions for F#, DataReaders
System.Data.DataSetExtensions.dll adds the ability to treat a DataTable as an IEnumerable<DataRow> as well as unboxing the values of individual cells in a way that gracefully handles DBNull by supporting System.Nullable. For example, in C# we can get the value of an integer column that contains nulls, and specify that DBNull should default to zero with a very concise syntax:
var total = myDataTable.AsEnumerable()
.Select(row => row.Field<int?>("MyColumn") ?? 0)
.Sum();
There are two areas where DataSetExtensions are lacking, however. First, it doesn't support IDataReader and second, it doesn't support the F# option type. The following code does both - it allows an IDataReader to be treated as a seq<IDataRecord>, and it can unbox values from either a reader or a dataset, with support for F# options or System.Nullable. Combined with the option-coalescing operator in another answer, this allows for code such as the following when working with a DataReader:
let total =
myReader.AsSeq
|> Seq.map (fun row -> row.Field<int option>("MyColumn") |? 0)
|> Seq.sum
Perhaps a more idiomatic F# way of ignoring database nulls would be...
let total =
myReader.AsSeq
|> Seq.choose (fun row -> row.Field<int option>("MyColumn"))
|> Seq.sum
Further, the extension methods defined below are usable from both F# and from C#/VB.
open System
open System.Data
open System.Reflection
open System.Runtime.CompilerServices
open Microsoft.FSharp.Collections
/// Ported from System.Data.DatasetExtensions.dll to add support for the Option type.
[<AbstractClass; Sealed>]
type private UnboxT<'a> private () =
// This class generates a converter function based on the desired output type,
// and then re-uses the converter function forever. Because the class itself is generic,
// different output types get different cached converter functions.
static let referenceField (value:obj) =
if value = null || DBNull.Value.Equals(value) then
Unchecked.defaultof<'a>
else
unbox value
static let valueField (value:obj) =
if value = null || DBNull.Value.Equals(value) then
raise <| InvalidCastException("Null cannot be converted to " + typeof<'a>.Name)
else
unbox value
static let makeConverter (target:Type) methodName =
Delegate.CreateDelegate(typeof<Converter<obj,'a>>,
typeof<UnboxT<'a>>
.GetMethod(methodName, BindingFlags.NonPublic ||| BindingFlags.Static)
.MakeGenericMethod([| target.GetGenericArguments().[0] |]))
|> unbox<Converter<obj,'a>>
|> FSharpFunc.FromConverter
static let unboxFn =
let theType = typeof<'a>
if theType.IsGenericType && not theType.IsGenericTypeDefinition then
let genericType = theType.GetGenericTypeDefinition()
if typedefof<Nullable<_>> = genericType then
makeConverter theType "NullableField"
elif typedefof<option<_>> = genericType then
makeConverter theType "OptionField"
else
invalidOp "The only generic types supported are Option<T> and Nullable<T>."
elif theType.IsValueType then
valueField
else
referenceField
static member private NullableField<'b when 'b : struct and 'b :> ValueType and 'b:(new:unit -> 'b)> (value:obj) =
if value = null || DBNull.Value.Equals(value) then
Nullable<_>()
else
Nullable<_>(unbox<'b> value)
static member private OptionField<'b> (value:obj) =
if value = null || DBNull.Value.Equals(value) then
None
else
Some(unbox<'b> value)
static member inline Unbox =
unboxFn
/// F# data-related extension methods.
[<AutoOpen>]
module FsDataEx =
type System.Data.IDataReader with
/// Exposes a reader's current result set as seq<IDataRecord>.
/// Reader is closed when sequence is fully enumerated.
member this.AsSeq =
seq { use reader = this
while reader.Read() do yield reader :> IDataRecord }
/// Exposes all result sets in a reader as seq<seq<IDataRecord>>.
/// Reader is closed when sequence is fully enumerated.
member this.AsMultiSeq =
let rowSeq (reader:IDataReader) =
seq { while reader.Read() do yield reader :> IDataRecord }
seq {
use reader = this
yield rowSeq reader
while reader.NextResult() do
yield rowSeq reader
}
/// Populates a new DataSet with the contents of the reader. Closes the reader after completion.
member this.ToDataSet () =
use reader = this
let dataSet = new DataSet(RemotingFormat=SerializationFormat.Binary, EnforceConstraints=false)
dataSet.Load(reader, LoadOption.OverwriteChanges, [| "" |])
dataSet
type System.Data.IDataRecord with
/// Gets a value from the record by name.
/// DBNull and null are returned as the default value for the type.
/// Supports both nullable and option types.
member this.Field<'a> (fieldName:string) =
this.[fieldName] |> UnboxT<'a>.Unbox
/// Gets a value from the record by column index.
/// DBNull and null are returned as the default value for the type.
/// Supports both nullable and option types.
member this.Field<'a> (ordinal:int) =
this.GetValue(ordinal) |> UnboxT<'a>.Unbox
type System.Data.DataRow with
/// Identical to the Field method from DatasetExtensions, but supports the F# Option type.
member this.Field2<'a> (columnName:string) =
this.[columnName] |> UnboxT<'a>.Unbox
/// Identical to the Field method from DatasetExtensions, but supports the F# Option type.
member this.Field2<'a> (columnIndex:int) =
this.[columnIndex] |> UnboxT<'a>.Unbox
/// Identical to the Field method from DatasetExtensions, but supports the F# Option type.
member this.Field2<'a> (column:DataColumn) =
this.[column] |> UnboxT<'a>.Unbox
/// Identical to the Field method from DatasetExtensions, but supports the F# Option type.
member this.Field2<'a> (columnName:string, version:DataRowVersion) =
this.[columnName, version] |> UnboxT<'a>.Unbox
/// Identical to the Field method from DatasetExtensions, but supports the F# Option type.
member this.Field2<'a> (columnIndex:int, version:DataRowVersion) =
this.[columnIndex, version] |> UnboxT<'a>.Unbox
/// Identical to the Field method from DatasetExtensions, but supports the F# Option type.
member this.Field2<'a> (column:DataColumn, version:DataRowVersion) =
this.[column, version] |> UnboxT<'a>.Unbox
/// C# data-related extension methods.
[<Extension; AbstractClass; Sealed>]
type CsDataEx private () =
/// Populates a new DataSet with the contents of the reader. Closes the reader after completion.
[<Extension>]
static member ToDataSet(this:IDataReader) =
this.ToDataSet()
/// Exposes a reader's current result set as IEnumerable{IDataRecord}.
/// Reader is closed when sequence is fully enumerated.
[<Extension>]
static member AsEnumerable(this:IDataReader) =
this.AsSeq
/// Exposes all result sets in a reader as IEnumerable{IEnumerable{IDataRecord}}.
/// Reader is closed when sequence is fully enumerated.
[<Extension>]
static member AsMultipleEnumerable(this:IDataReader) =
this.AsMultiSeq
/// Gets a value from the record by name.
/// DBNull and null are returned as the default value for the type.
/// Supports both nullable and option types.
[<Extension>]
static member Field<'T> (this:IDataRecord, fieldName:string) =
this.Field<'T>(fieldName)
/// Gets a value from the record by column index.
/// DBNull and null are returned as the default value for the type.
/// Supports both nullable and option types.
[<Extension>]
static member Field<'T> (this:IDataRecord, ordinal:int) =
this.Field<'T>(ordinal)
Handling arguments in a command line application:
//We assume that the actual meat is already defined in function
// DoStuff (string -> string -> string -> unit)
let defaultOutOption = "N"
let defaultUsageOption = "Y"
let usage =
"Scans a folder for and outputs results.\n" +
"Usage:\n\t MyApplication.exe FolderPath [IncludeSubfolders (Y/N) : default=" +
defaultUsageOption + "] [OutputToFile (Y/N): default=" + defaultOutOption + "]"
let HandlArgs arr =
match arr with
| [|d;u;o|] -> DoStuff d u o
| [|d;u|] -> DoStuff d u defaultOutOption
| [|d|] -> DoStuff d defaultUsageOption defaultOutOption
| _ ->
printf "%s" usage
Console.ReadLine() |> ignore
[<EntryPoint>]
let main (args : string array) =
args |> HandlArgs
0
(I had a vague memory of this technique being inspired by Robert Pickering, but can't find a reference now)
A handy cache function that keeps up to max (key,reader(key)) in a dictionary and use a SortedList to track the MRU keys
let Cache (reader: 'key -> 'value) max =
let cache = new Dictionary<'key,LinkedListNode<'key * 'value>>()
let keys = new LinkedList<'key * 'value>()
fun (key : 'key) -> (
let found, value = cache.TryGetValue key
match found with
|true ->
keys.Remove value
keys.AddFirst value |> ignore
(snd value.Value)
|false ->
let newValue = key,reader key
let node = keys.AddFirst newValue
cache.[key] <- node
if (keys.Count > max) then
let lastNode = keys.Last
cache.Remove (fst lastNode.Value) |> ignore
keys.RemoveLast() |> ignore
(snd newValue))
Creating XElements
Nothing amazing, but I keep getting caught out by the implicit conversion of XNames:
#r "System.Xml.Linq.dll"
open System.Xml.Linq
//No! ("type string not compatible with XName")
//let el = new XElement("MyElement", "text")
//better
let xn s = XName.op_Implicit s
let el = new XElement(xn "MyElement", "text")
//or even
let xEl s o = new XElement(xn s, o)
let el = xEl "MyElement" "text"
Pairwise and pairs
I always expect Seq.pairwise to give me [(1,2);(3;4)] and not [(1,2);(2,3);(3,4)]. Given that neither exist in List, and that I needed both, here's the code for future reference. I think they're tail recursive.
//converts to 'windowed tuples' ([1;2;3;4;5] -> [(1,2);(2,3);(3,4);(4,5)])
let pairwise lst =
let rec loop prev rem acc =
match rem with
| hd::tl -> loop hd tl ((prev,hd)::acc)
| _ -> List.rev acc
loop (List.head lst) (List.tail lst) []
//converts to 'paged tuples' ([1;2;3;4;5;6] -> [(1,2);(3,4);(5,6)])
let pairs lst =
let rec loop rem acc =
match rem with
| l::r::tl -> loop tl ((l,r)::acc)
| l::[] -> failwith "odd-numbered list"
| _ -> List.rev acc
loop lst []
Naive CSV reader (i.e., won't handle anything nasty)
(Using filereadlines and List.transpose from other answers here)
///Given a file path, returns a List of row lists
let ReadCSV =
filereadlines
>> Array.map ( fun line -> line.Split([|',';';'|]) |> List.ofArray )
>> Array.toList
///takes list of col ids and list of rows,
/// returns array of columns (in requested order)
let GetColumns cols rows =
//Create filter
let pick cols (row:list<'a>) = List.map (fun i -> row.[i]) cols
rows
|> transpose //change list of rows to list of columns
|> pick cols //pick out the columns we want
|> Array.ofList //an array output is easier to index for user
Example
"C:\MySampleCSV"
|> ReadCSV
|> List.tail //skip header line
|> GetColumns [0;3;1] //reorder columns as well, if needs be.
Date Range
simple but useful list of dates between fromDate and toDate
let getDateRange fromDate toDate =
let rec dates (fromDate:System.DateTime) (toDate:System.DateTime) =
seq {
if fromDate <= toDate then
yield fromDate
yield! dates (fromDate.AddDays(1.0)) toDate
}
dates fromDate toDate
|> List.ofSeq
toggle code to sql
More trivial than most on this list, but handy nonetheless:
I'm always taking sql in and out of code to move it to a sql environment during development. Example:
let sql = "select a,b,c "
+ "from table "
+ "where a = 1"
needs to be 'stripped' to:
select a,b,c
from table
where a = 1
keeping the formatting. It's a pain to strip out the code symbols for the sql editor, then put them back again by hand when I've got the sql worked out. These two functions toggle the sql back and forth from code to stripped:
// reads the file with the code quoted sql, strips code symbols, dumps to FSI
let stripForSql fileName =
File.ReadAllText(fileName)
|> (fun s -> Regex.Replace(s, "\+(\s*)\"", ""))
|> (fun s -> s.Replace("\"", ""))
|> (fun s -> Regex.Replace(s, ";$", "")) // end of line semicolons
|> (fun s -> Regex.Replace(s, "//.+", "")) // get rid of any comments
|> (fun s -> printfn "%s" s)
then when you are ready to put it back into your code source file:
let prepFromSql fileName =
File.ReadAllText(fileName)
|> (fun s -> Regex.Replace(s, #"\r\n", " \"\r\n+\"")) // matches newline
|> (fun s -> Regex.Replace(s, #"\A", " \""))
|> (fun s -> Regex.Replace(s, #"\z", " \""))
|> (fun s -> printfn "%s" s)
I'd love to get rid of the input file but can't even begin to grok how to make that happen. anyone?
edit:
I figured out how to eliminate the requirement of a file for these functions by adding a windows forms dialog input/output. Too much code to show, but for those who would like to do such a thing, that's how I solved it.
Pascal's Triangle (hey, someone might find it useful)
So we want to create a something like this:
1
1 1
1 2 1
1 3 3 1
1 4 6 4 1
Easy enough:
let rec next = function
| [] -> []
| x::y::xs -> (x + y)::next (y::xs)
| x::xs -> x::next xs
let pascal n =
seq { 1 .. n }
|> List.scan (fun acc _ -> next (0::acc) ) [1]
The next function returns a new list where each item[i] = item[i] + item[i + 1].
Here's the output in fsi:
> pascal 10 |> Seq.iter (printfn "%A");;
[1]
[1; 1]
[1; 2; 1]
[1; 3; 3; 1]
[1; 4; 6; 4; 1]
[1; 5; 10; 10; 5; 1]
[1; 6; 15; 20; 15; 6; 1]
[1; 7; 21; 35; 35; 21; 7; 1]
[1; 8; 28; 56; 70; 56; 28; 8; 1]
[1; 9; 36; 84; 126; 126; 84; 36; 9; 1]
[1; 10; 45; 120; 210; 252; 210; 120; 45; 10; 1]
For the adventurous, here's a tail-recursive version:
let rec next2 cont = function
| [] -> cont []
| x::y::xs -> next2 (fun l -> cont <| (x + y)::l ) <| y::xs
| x::xs -> next2 (fun l -> cont <| x::l ) <| xs
let pascal2 n =
set { 1 .. n }
|> Seq.scan (fun acc _ -> next2 id <| 0::acc)) [1]
Flatten a List
if you have something like this:
let listList = [[1;2;3;];[4;5;6]]
and want to 'flatten' it down to a singe list so the result is like this:
[1;2;3;4;5;6]
it can be done thusly:
let flatten (l: 'a list list) =
seq {
yield List.head (List.head l)
for a in l do yield! (Seq.skip 1 a)
}
|> List.ofSeq
List comprehensions for float
This [23.0 .. 1.0 .. 40.0] was marked as deprecated a few versions backed.
But apparently, this works:
let dl = 9.5 / 11.
let min = 21.5 + dl
let max = 40.5 - dl
let a = [ for z in min .. dl .. max -> z ]
let b = a.Length
(BTW, there's a floating point gotcha in there. Discovered at fssnip - the other place for F# snippets)
Parallel map
let pmap f s =
seq { for a in s -> async { return f s } }
|> Async.Parallel
|> Async.Run