Apologies for a rookie question. I'm trying to change my mental paradigm from procedural to functional.
For instance, suppose I have a list of names that I want to print like this "John, Paul, George, and Ringo." But this code does not satisfy:
let names = [ "John"; "Paul"; "George"; "Ringo" ]
names |> Seq.iter (fun s -> printf "%s, " s)
My procedural instinct is to seek a way to insinuate a predicate into that lambda so that it can branch between ", " or ", and " or ". " depending upon where we're at iterating the sequence. I think that's wrong, but I'm feeling around for what's right.
Would it be better to split the sequence in parts?
In this case it seems that we want to split the sequence into parts corresponding to distinct delimiter behaviors. We want to split it at the end, so we can't use Seq. But we can use List.splitAt instead.
let start, ending = List.splitAt (names.Length - 1) names
let penultimate, last = List.splitAt 1 ending
start |> Seq.iter (fun s -> printf "%s, " s)
penultimate |> Seq.iter (fun s -> printf "%s, and " s)
last |> Seq.iter (fun s -> printf "%s. " s)
Is this a righteous approach? Is there a better solution I've overlooked? Am I thinking along the right lines?
The general approach I take to tackle these kind of problems is to split them into smaller parts and solve individually:
an empty list [] results in ""
one element ["a"] results in "a."
two elements [ "a"; "b" ] result in "a and b."
more elements (that is a :: rest) result in "a, " + takeCareOf rest, where takeCareOf follows above rules. Note that we don't need to know the length of the full list.
Above recipe directly translates to F# (and functional languages in general):
let rec commaAndDot' = function
| [] -> ()
| [ a ] -> printfn "%s." a
| a :: [ b ] -> printfn "%s and %s." a b
| a :: rest -> printf "%s, " a; commaAndDot' rest
Are we done yet? No, commaAndDot' violates the Single Responsibility Principle because the function implements our 'business logic' and prints to the console. Let's fix that:
let rec commaAndDot'' = function
| [] -> ""
| [ a ] -> sprintf "%s." a
| a :: [ b ] -> sprintf "%s and %s." a b
| a :: rest -> sprintf "%s, " a + commaAndDot'' rest
As an additional benefit we can now call the function in parallel and the output does not get mixed up anymore.
Are we done yet? No, above function is not tail-recursive (we need to compute commaAndDot'' rest before concatenating it to the current result) and would blow the stack for large lists. A standard approach to fixing this is to introduce an accumulator acc:
let commaAndDot''' words =
let rec helper acc = function
| [] -> acc
| [ a ] -> sprintf "%s%s." acc a
| a :: [ b ] -> sprintf "%s%s and %s." acc a b
| a :: rest -> helper (acc + sprintf "%s, " a) rest
helper "" words
Are we done yet? No, commaAndDot''' creates a lot of strings for intermediate results. Thanks to F# not being a pure language, we can leverage local (private, non-observable) mutation to optimize for memory and speed:
let commaAndDot words =
let sb = System.Text.StringBuilder()
let rec helper = function
| [] -> sb
| [ a ] -> sprintf "%s." a |> sb.Append
| a :: [ b ] -> sprintf "%s and %s." a b |> sb.Append
| a :: rest ->
sprintf "%s, " a |> sb.Append |> ignore
helper rest
helper words |> string
Are we done yet? Probably... at least this is something I would consider idiomatic F# and happily commit. For optimising further (e.g. Appending commas and dots separately or changing the order of the patterns) I'd first write micro-benchmarks before sacrificing readability.
All versions generate the same output:
commaAndDot [] // ""
commaAndDot [ "foo" ] // "foo."
commaAndDot [ "foo"; "bar" ] // "foo and bar."
commaAndDot [ "Hello"; "World"; "F#" ] // "Hello, World and F#."
Update: SCNR, created a benchmark... results are below as a HTML snippet (for nice tabular data).
BuilderOpt is the StringBuilder version with the [] case moved to the bottom,
BuilderChained is with chained Append calls, e.g. sb.Append(a).Append(" and ").Append(b) and BuilderFormat is e.g. sb.AppendFormat("{0} and {1}", a, b). Full source code available.
As expected, 'simpler' versions perform better for small lists, the larger the list the better BuilderChained. Concat performs better than I expected but does not produce the right output (missing ".", lacking one case). Yield gets rather slow...
<!DOCTYPE html>
<html lang='en'>
<head>
<meta charset='utf-8' />
<title>Benchmark.CommaAndDot</title>
<style type="text/css">
table { border-collapse: collapse; display: block; width: 100%; overflow: auto; }
td, th { padding: 6px 13px; border: 1px solid #ddd; }
tr { background-color: #fff; border-top: 1px solid #ccc; }
tr:nth-child(even) { background: #f8f8f8; }
</style>
</head>
<body>
<pre><code>
BenchmarkDotNet=v0.11.1, OS=Windows 10.0.16299.726 (1709/FallCreatorsUpdate/Redstone3)
Intel Core i7 CPU 950 3.07GHz (Nehalem), 1 CPU, 8 logical and 4 physical cores
Frequency=2998521 Hz, Resolution=333.4977 ns, Timer=TSC
[Host] : .NET Framework 4.7.2 (CLR 4.0.30319.42000), 64bit LegacyJIT-v4.7.3190.0 DEBUG
DefaultJob : .NET Framework 4.7.2 (CLR 4.0.30319.42000), 64bit RyuJIT-v4.7.3190.0
</code></pre>
<pre><code></code></pre>
<table>
<thead><tr><th> Method</th><th>Verbosity</th><th> Mean</th><th>Error</th><th>StdDev</th><th> Median</th><th>Scaled</th><th>ScaledSD</th>
</tr>
</thead><tbody><tr><td>Concat</td><td>0</td><td>39.905 ns</td><td>0.0592 ns</td><td>0.0494 ns</td><td>39.906 ns</td><td>1.02</td><td>0.11</td>
</tr><tr><td>Yield</td><td>0</td><td>27.235 ns</td><td>0.0772 ns</td><td>0.0603 ns</td><td>27.227 ns</td><td>0.69</td><td>0.07</td>
</tr><tr><td>Accumulator</td><td>0</td><td>1.956 ns</td><td>0.0109 ns</td><td>0.0096 ns</td><td>1.954 ns</td><td>0.05</td><td>0.01</td>
</tr><tr><td>Builder</td><td>0</td><td>32.384 ns</td><td>0.2986 ns</td><td>0.2331 ns</td><td>32.317 ns</td><td>0.82</td><td>0.09</td>
</tr><tr><td>BuilderOpt</td><td>0</td><td>33.664 ns</td><td>1.0371 ns</td><td>0.9194 ns</td><td>33.402 ns</td><td>0.86</td><td>0.09</td>
</tr><tr><td>BuilderChained</td><td>0</td><td>39.671 ns</td><td>1.2097 ns</td><td>3.5669 ns</td><td>41.339 ns</td><td>1.00</td><td>0.00</td>
</tr><tr><td>BuilderFormat</td><td>0</td><td>40.276 ns</td><td>0.8909 ns</td><td>1.8792 ns</td><td>39.494 ns</td><td>1.02</td><td>0.12</td>
</tr><tr><td>Concat</td><td>1</td><td>153.116 ns</td><td>1.1592 ns</td><td>0.9050 ns</td><td>152.706 ns</td><td>0.87</td><td>0.01</td>
</tr><tr><td>Yield</td><td>1</td><td>154.522 ns</td><td>0.2890 ns</td><td>0.2256 ns</td><td>154.479 ns</td><td>0.88</td><td>0.00</td>
</tr><tr><td>Accumulator</td><td>1</td><td>223.342 ns</td><td>0.3678 ns</td><td>0.2872 ns</td><td>223.412 ns</td><td>1.27</td><td>0.00</td>
</tr><tr><td>Builder</td><td>1</td><td>232.194 ns</td><td>0.2951 ns</td><td>0.2465 ns</td><td>232.265 ns</td><td>1.32</td><td>0.00</td>
</tr><tr><td>BuilderOpt</td><td>1</td><td>232.016 ns</td><td>0.5654 ns</td><td>0.4722 ns</td><td>232.170 ns</td><td>1.31</td><td>0.00</td>
</tr><tr><td>BuilderChained</td><td>1</td><td>176.473 ns</td><td>0.3918 ns</td><td>0.3272 ns</td><td>176.341 ns</td><td>1.00</td><td>0.00</td>
</tr><tr><td>BuilderFormat</td><td>1</td><td>219.262 ns</td><td>6.7995 ns</td><td>6.3603 ns</td><td>217.003 ns</td><td>1.24</td><td>0.03</td>
</tr><tr><td>Concat</td><td>10</td><td>1,284.042 ns</td><td>1.7035 ns</td><td>1.4225 ns</td><td>1,283.443 ns</td><td>1.68</td><td>0.05</td>
</tr><tr><td>Yield</td><td>10</td><td>6,532.667 ns</td><td>12.6169 ns</td><td>10.5357 ns</td><td>6,533.504 ns</td><td>8.55</td><td>0.24</td>
</tr><tr><td>Accumulator</td><td>10</td><td>2,701.483 ns</td><td>4.8509 ns</td><td>4.5376 ns</td><td>2,700.208 ns</td><td>3.54</td><td>0.10</td>
</tr><tr><td>Builder</td><td>10</td><td>1,865.668 ns</td><td>5.0275 ns</td><td>3.9252 ns</td><td>1,866.920 ns</td><td>2.44</td><td>0.07</td>
</tr><tr><td>BuilderOpt</td><td>10</td><td>1,820.402 ns</td><td>2.7853 ns</td><td>2.3258 ns</td><td>1,820.464 ns</td><td>2.38</td><td>0.07</td>
</tr><tr><td>BuilderChained</td><td>10</td><td>764.334 ns</td><td>19.8528 ns</td><td>23.6334 ns</td><td>756.988 ns</td><td>1.00</td><td>0.00</td>
</tr><tr><td>BuilderFormat</td><td>10</td><td>1,177.186 ns</td><td>1.9584 ns</td><td>1.6354 ns</td><td>1,177.897 ns</td><td>1.54</td><td>0.04</td>
</tr><tr><td>Concat</td><td>100</td><td>25,579.773 ns</td><td>824.1504 ns</td><td>688.2028 ns</td><td>25,288.873 ns</td><td>5.33</td><td>0.14</td>
</tr><tr><td>Yield</td><td>100</td><td>421,872.560 ns</td><td>902.5023 ns</td><td>753.6302 ns</td><td>421,782.071 ns</td><td>87.87</td><td>0.23</td>
</tr><tr><td>Accumulator</td><td>100</td><td>80,579.168 ns</td><td>227.7392 ns</td><td>177.8038 ns</td><td>80,547.868 ns</td><td>16.78</td><td>0.05</td>
</tr><tr><td>Builder</td><td>100</td><td>15,047.790 ns</td><td>26.2248 ns</td><td>21.8989 ns</td><td>15,048.903 ns</td><td>3.13</td><td>0.01</td>
</tr><tr><td>BuilderOpt</td><td>100</td><td>15,287.117 ns</td><td>39.8679 ns</td><td>31.1262 ns</td><td>15,293.739 ns</td><td>3.18</td><td>0.01</td>
</tr><tr><td>BuilderChained</td><td>100</td><td>4,800.966 ns</td><td>11.3614 ns</td><td>10.0716 ns</td><td>4,801.450 ns</td><td>1.00</td><td>0.00</td>
</tr><tr><td>BuilderFormat</td><td>100</td><td>8,382.896 ns</td><td>87.8963 ns</td><td>68.6236 ns</td><td>8,368.400 ns</td><td>1.75</td><td>0.01</td>
</tr></tbody></table>
</body>
</html>
I prefer using String.concat:
let names = [ "John"; "Paul"; "George"; "Ringo" ]
names
|> List.mapi (fun i n -> if i = names.Length - 1 && i > 0 then "and " + n else n)
|> String.concat ", "
|> printfn "%s"
Basic techniques are mentioned in the accepted answer: problem deconstruction and separation of concerns. There is either no element, or there is an element followed by either ., , and, or ,, depending on its position relative to the end of the input sequence.
Assuming that the input is of type string list, this can be fairly well expressed by a recursive, pattern matching function definition, wrapped inside a list sequence expression to ensure tail recursion. The match does nothing if the input is empty, so it returns an empty list; it returns a sub-list for the other terminating case, otherwise it appends to the sub-list the results of the recursion.
The concatenation as the desired target type string is a separate, final step, as proposed in another answer.
let rec seriesComma xs = [
match xs with
| [] -> ()
| [x] -> yield! [x; "."]
| x::[y] -> yield! [x; ", and "]; yield! seriesComma [y]
| x::xs -> yield! [x; ", "]; yield! seriesComma xs ]
["Chico"; "Harpo"; "Groucho"; "Gummo"; "Zeppo"]
|> seriesComma |> String.concat ""
// val it : string = "Chico, Harpo, Groucho, Gummo, and Zeppo."
Seq.Reduce is the simplest way to make a delimited list, but including the "and" before the last item adds some complexity. Below I show a way to do it in two steps, but the recursive approach in the accepted answer is probably more true to the Functional Programming paradigm.
let names = [ "John"; "Paul"; "George"; "Ringo" ]
let delimitedNames = names |> Seq.reduce (fun x y -> sprintf "%s, %s" x y)
let replaceLastOccurrence (hayStack: string) (needle: string) (newNeedle: string) =
let idx = hayStack.LastIndexOf needle
match idx with
| -1 -> hayStack
| _ -> hayStack.Remove(idx, needle.Length).Insert(idx, newNeedle)
replaceLastOccurrence delimitedNames "," ", and"
See https://msdn.microsoft.com/en-us/visualfsharpdocs/conceptual/seq.reduce%5B%27t%5D-function-%5Bfsharp%5D?f=255&MSPPError=-2147217396
Well, a more functional-looking solution could be something like this:
let names = [ "John"; "Paul"; "George"; "Ringo" ]
names
|> Seq.tailBack
|> Seq.iter (fun s -> printf "%s, " s)
names
|> Seq.last
|> fun s -> printf "and %s" s
Where tailBack can be defined in some SequenceExtensions.fs like
module Seq
let tailBack seq =
seq
|> Seq.rev
|> Seq.tail
|> Seq.rev
This way you do not deal much with indexes, variables and all that procedural stuff.
Ideally you would leverage options here, like
names
|> Seq.tryLast
|> Option.iter (fun s -> printf "and %s" s)
With this you would also avoid possible argument exceptions. But options in functional programming is another (nice) concept than sequences.
Also, here a particular task matters. I believe this solution is quite inefficient - we iterate the sequence too many times. Maybe in some cases fussing with indexes will be the way to go.
Related
I have a problem with how nested functions should be implemented in OCaml, i need the output (list) of one function to be the input of another. And both should be recursive. The problem is i've played around with the parameters and they arent feeding properly:
let toComb sentence =
let rec listCleanup sentence =
match sentence with
| [] -> []
| h::t when h = "" -> listCleanup t
| h::t -> h::listCleanup t
in
let rec toString listCleanup sentence =
match listCleanup sentence with
| [] -> ""
| [element] -> element
| h::t -> h ^ " " ^ toString listCleanup sentence
in
toString listCleanup sentence;;
If I use the function and its parameter as a parameter, there's a stack overflow, but if I use just the function without a parameter, I get a mismatch of parameters. What should be the fix here?
To correct your code, here is what would work properly:
let to_comb sentence =
let rec cleanup s = match s with
| [] -> []
| ""::tail -> cleanup tail
| hd::tail -> hd::cleanup tail in
let rec to_string s = match s with
| [] -> ""
| [x] -> x
| hd::tail -> hd ^ " " ^ to_string tail in
to_string (cleanup s)
Note that I only call cleanup once, because you only ever need to clean the whole sequence only once. However, turns out both of these function can be expressed more simply with predefined OCaml function:
let to_comb sentence =
sentence
|> List.filter (fun s -> s <> "")
|> String.concat " "
You could almost read this code out loud to get a description of what it does. It starts with a sentence, filters the empty words in it, then concatenates them with spaces in between.
I am a new at F# and i try to do this task:
Make a function compare : string list -> string list -> int that takes two string lists and returns: -1, 0 or 1
Please help. I spend a lot of time, and i can not understand how to implement this task.
Given the task I assume what your professor wants to teach you with this exercise. I'll try to give you a starting point without
Confusing you
Presenting a 'done-deal' solution
I assume the goal of this task is to work with recursive functions and pattern matching to element-wise compare their elements. It could looks somewhat like this here
open System
let aList = [ "Apple"; "Banana"; "Coconut" ]
let bList = [ "Apple"; "Banana"; "Coconut" ]
let cList = [ "Apple"; "Zebra" ]
let rec doSomething f (a : string list) (b : string list) =
match (a, b) with
| ([], []) ->
printfn "Both are empty"
| (x::xs, []) ->
printfn "A has elements (we can unpack the first element as x and the rest as xs) and B is empty"
| ([], x::xs) ->
printfn "A is empty and B has elements (we can unpack the first element as x and the rest as xs)"
| (x::xs, y::ys) ->
f x y
printfn "Both A and B have elements. We can unpack them as the first elements x and y and their respective tails xs and ys"
doSomething f xs ys
let isItTheSame (a : string) (b : string) =
if String.Equals(a, b) then
printfn "%s is equals to %s" a b
else
printfn "%s is not equals to %s" a b
doSomething isItTheSame aList bList
doSomething isItTheSame aList cList
The example has three different lists, two of them being equal and one of them being different. The doSomething function takes a function (string -> string -> unit) and two lists of strings.
Within the function you see a pattern match as well as a recursive call of doSomething in the last match block. The signatures aren't exactly what you need and you might want to think about how to change the parametrization for cases where you don't want to stop the recursion (the last match block - if the strings are equal you want to keep on comparing, right?).
Just take the code and try it out in FSI. I'm confident, that you'll find the solution 🙂
In F# many collections are comparable if their element type is:
let s1 = [ "a"; "b" ]
let s2 = [ "foo"; "bar" ]
compare s1 s2 // -5
let f1 = [ (fun () -> 1); fun () -> 2 ]
let f2 = [ (fun () -> 3); fun () -> 42 ]
// compare f1 f2 (* error FS0001: The type '(unit -> int)' does not support the 'comparison' constraint. *)
so
let slcomp (s1 : string list) s2 = compare s1 s2 |> sign
Posting for reference as the original question is answered already.
I'm currently generating a sequence in a similar way to:
migrators
|> Seq.map (fun m -> m())
The migrator function is ultimately returning a discriminated union like:
type MigratorResult =
| Success of string * TimeSpan
| Error of string * Exception
I want to stop the map once I encounter my first Error but I need to include the Error in the final sequence.
I have something like the following to display a final message to the user
match results |> List.rev with
| [] -> "No results equals no migrators"
| head :: _ ->
match head with
| Success (dt, t) -> "All migrators succeeded"
| Error (dt, ex) -> "Migration halted owing to error"
So I need:
A way to stop the mapping when one of the map steps produces an Error
A way to have that error be the final element added to the sequence
I appreciate there may be a different sequence method other than map that will do this, I'm new to F# and searching online hasn't yielded anything as yet!
I guess there are multiple approaches here, but one way would be to use unfold:
migrators
|> Seq.unfold (fun ms ->
match ms with
| m :: tl ->
match m () with
| Success res -> Some (Success res, tl)
| Error res -> Some (Error res, [])
| [] -> None)
|> List.ofSeq
Note the List.ofSeq at the end, that's just there for realizing the sequence. A different way to go would be to use sequence comprehensions, some might say it results in a clearer code.
The ugly things Tomaš alludes to are 1) mutable state, and 2) manipulation of the underlying enumerator. A higher-order function which returns up to and including when the predicate holds would then look like this:
module Seq =
let takeUntil pred (xs : _ seq) = seq{
use en = xs.GetEnumerator()
let flag = ref true
while !flag && en.MoveNext() do
flag := not <| pred en.Current
yield en.Current }
seq{1..10} |> Seq.takeUntil (fun x -> x % 5 = 0)
|> Seq.toList
// val it : int list = [1; 2; 3; 4; 5]
For your specific application, you'd map the cases of the DU to a boolean.
(migrators : seq<MigratorResult>)
|> Seq.takeUntil (function Success _ -> false | Error _ -> true)
I think the answer from #scrwtp is probably the nicest way to do this if your input is reasonably small (and you can turn it into an F# list to use pattern matching). I'll add one more version, which works when your input is just a sequence and you do not want to turn it into a list.
Essentially, you want to do something that's almost like Seq.takeWhile, but it gives you one additional item at the end (the one, for which the predicate fails).
To use a simpler example, the following returns all numbers from a sequence until one that is divisible by 5:
let nums = [ 2 .. 10 ]
nums
|> Seq.map (fun m -> m % 5)
|> Seq.takeWhile (fun n -> n <> 0)
So, you basically just need to look one element ahead - to do this, you could use Seq.pairwise which gives you the current and the next element in the sequence"
nums
|> Seq.map (fun m -> m % 5)
|> Seq.pairwise // Get sequence of pairs with the next value
|> Seq.takeWhile (fun (p, n) -> p <> 0) // Look at the next value for test
|> Seq.mapi (fun i (p, n) -> // For the first item, we return both
if i = 0 then [p;n] else [n]) // for all other, we return the second
|> Seq.concat
The only ugly thing here is that you then need to flatten the sequence again using mapi and concat.
This is not very nice, so a good thing to do would be to define your own higher-order function like Seq.takeUntilAfter that encapsulates the behavior you need (and hides all the ugly things). Then your code could just use the function and look nice & readable (and you can experiment with other ways of implementing this).
I am currently doing reasonably well in functional programming using F#. I tend, however, to do a lot of programming using recursion, when it seems that there are better idioms in the F#/functional programming community. So in the spirit of learning, is there a better/more idiomatic way of writing the function below without recursion?
let rec convert line =
if line.[0..1] = " " then
match convert line.[2..] with
| (i, subline) -> (i+1, subline)
else
(0, line)
with results such as:
> convert "asdf";;
val it : int * string = (0, "asdf")
> convert " asdf";;
val it : int * string = (1, "asdf")
> convert " asdf";;
val it : int * string = (3, "asdf")
Recursion is the basic mechanism for writing loops in functional languages, so if you need to iterate over characters (as you do in your sample), then recursion is what you need.
If you want to improve your code, then you should probably avoid using line.[2..] because that is going to be inefficient (strings are not designed for this kind of processing). It is better to convert the string to a list and then process it:
let convert (line:string) =
let rec loop acc line =
match line with
| ' '::' '::rest -> loop (acc + 1) rest
| _ -> (acc, line)
loop 0 (List.ofSeq line)
You can use various functions from the standard library to implement this in a more shorter way, but they are usually recursive too (you just do not see the recursion!), so I think using functions like Seq.unfold and Seq.fold is still recursive (and it looks way more complex than your code).
A more concise approach using standard libraries is to use the TrimLeft method (see comments), or using standard F# library functions, do something like this:
let convert (line:string) =
// Count the number of spaces at the beginning
let spaces = line |> Seq.takeWhile (fun c -> c = ' ') |> Seq.length
// Divide by two - we want to count & skip two-spaces only
let count = spaces / 2
// Get substring starting after all removed two-spaces
count, line.[(count * 2) ..]
EDIT Regarding the performance of string vs. list processing, the problem is that slicing allocates a new string (because that is how strings are represented on the .NET platform), while slicing a list just changes a reference. Here is a simple test:
let rec countList n s =
match s with
| x::xs -> countList (n + 1) xs
| _ -> n
let rec countString n (s:string) =
if s.Length = 0 then n
else countString (n + 1) (s.[1 ..])
let l = [ for i in 1 .. 10000 -> 'x' ]
let s = new System.String('x', 10000)
#time
for i in 0 .. 100 do countList 0 l |> ignore // 0.002 sec (on my machine)
for i in 0 .. 100 do countString 0 s |> ignore // 5.720 sec (on my machine)
Because you traverse the string in a non-uniform way, a recursive solution is much more suitable in this example. I would rewrite your tail-recursive solution for readability as follows:
let convert (line: string) =
let rec loop i line =
match line.[0..1] with
| " " -> loop (i+1) line.[2..]
| _ -> i, line
loop 0 line
Since you asked, here is a (bizarre) non-recursive solution :).
let convert (line: string) =
(0, line) |> Seq.unfold (fun (i, line) ->
let subline = line.[2..]
match line.[0..1] with
| " " -> Some((i+1, subline), (i+1, subline))
| _ -> None)
|> Seq.fold (fun _ x -> x) (0, line)
Using tail recursion, it can be written as
let rec convert_ acc line =
if line.[0..1] <> " " then
(acc, line)
else
convert_ (acc + 1) line.[2..]
let convert = convert_ 0
still looking for a non-recursive answer, though.
Here's a faster way to write your function -- it checks the characters explicitly instead of using string slicing (which, as Tomas said, is slow); it's also tail-recursive. Finally, it uses a StringBuilder to create the "filtered" string, which will provide better performance once your input string reaches a decent length (though it'd be a bit slower for very small strings due to the overhead of creating the StringBuilder).
let convert' str =
let strLen = String.length str
let sb = System.Text.StringBuilder strLen
let rec convertRec (count, idx) =
match strLen - idx with
| 0 ->
count, sb.ToString ()
| 1 ->
// Append the last character in the string to the StringBuilder.
sb.Append str.[idx] |> ignore
convertRec (count, idx + 1)
| _ ->
if str.[idx] = ' ' && str.[idx + 1] = ' ' then
convertRec (count + 1, idx + 2)
else
sb.Append str.[idx] |> ignore
convertRec (count, idx + 1)
// Call the internal, recursive implementation.
convertRec (0, 0)
Take this example code (ignore it being horribly inefficient for the moment)
let listToString (lst:list<'a>) = ;;' prettify fix
let rec inner (lst:list<'a>) buffer = ;;' prettify fix
match List.length lst with
| 0 -> buffer
| _ -> inner (List.tl lst) (buffer + ((List.hd lst).ToString()))
inner lst ""
This is a common pattern I keep coming across in F#, I need to have an inner function who recurses itself over some value - and I only need this function once, is there in any way possible to call a lambda from within it self (some magic keyword or something) ? I would like the code to look something like this:
let listToString2 (lst:list<'a>) = ;;' prettify fix
( fun
(lst:list<'a>) buffer -> match List.length lst with ;;' prettify fix
| 0 -> buffer
| _ -> ##RECURSE## (List.tl lst) (buffer + ((List.hd lst).ToString()))
) lst ""
But as you might expect there is no way to refer to the anonymous function within itself, which is needed where I put ##RECURSE##
Yes, it's possible using so called y-combinators (or fixed-point combinators). Ex:
let rec fix f x = f (fix f) x
let fact f = function
| 0 -> 1
| x -> x * f (x-1)
let _ = (fix fact) 5 (* evaluates to "120" *)
I don't know articles for F# but this haskell entry might also be helpful.
But: I wouldn't use them if there is any alternative - They're quite hard to understand.
Your code (omit the type annotations here) is a standard construct and much more expressive.
let listToString lst =
let rec loop acc = function
| [] -> acc
| x::xs -> loop (acc ^ (string x)) xs
loop "" lst
Note that although you say you use the function only once, technically you refer to it by name twice, which is why it makes sense to give it a name.