I'm starting to like the Swift string formatting since it uses variable names in the string rather than ambiguous formatting tags like "%#"
I want to load a large string from a file that has Swift-style formatting in it (like this)
Now is the time for all good \(who) to come to babble incoherently.
Then I want to feed the contents of that String variable into a statement that lest me replace
\(who)
with the contents of the constant/variable who at runtime.
The code below works with a string constant as the formatting string.
let who = "programmers"
let aString = "Now is the time for all good \(who) to come to babble incoherently."
That code does formatting of a quoted string that appears in-line in my code.
Instead I want something like the code
let formatString = "Now is the time for all good %# to come to babble incoherently."
aString = String(format: formatString, who)
But where I can pass in a Swift-style format string in a constant/variable I read from a file.
Is that possible? I didn't have any luck searching for it since I wasn't exactly sure what search terms to use.
I can always use C-style string formatting and the String class' initWithFormat method if I have to...
I don't think there's a way to do this. String interpolation is implemented via conforming to the StringInterpolationConvertible protocol, and presumably you're hoping to tap into that in the same way you can tap into the methods required by StringLiteralConvertible, a la:
let someString = toString(42)
// this is the method String implements to conform to StringLiteralConvertible
let anotherString = String(stringLiteral: someString)
// anotherString will be "42"
print(anotherString)
Unfortunately, you can't do quite the same trick with StringInterpolationConvertible. Seeing how the protocol works may help:
struct MyString: Printable {
let actualString: String
var description: String { return actualString }
}
extension MyString: StringInterpolationConvertible {
// first, this will get called for each "segment"
init<T>(stringInterpolationSegment expr: T) {
println("Processing segment: " + toString(expr))
actualString = toString(expr)
}
// here is a type-specific override for Int, that coverts
// small numbers into words:
init(stringInterpolationSegment expr: Int) {
if (0..<4).contains(expr) {
println("Embigening \(expr)")
let numbers = ["zeo","one","two","three"]
actualString = numbers[expr]
}
else {
println("Processing segment: " + toString(expr))
actualString = toString(expr)
}
}
// finally, this gets called with an array of all of the
// converted segments
init(stringInterpolation strings: MyString...) {
// strings will be a bunch of MyString objects
actualString = "".join(strings.map { $0.actualString })
}
}
let number = 3
let aString: MyString = "Then shalt thou count to \(number), no more, no less."
println(aString)
// prints "Then shalt thou count to three, no more, no less."
So, while you can call String.init(stringInterpolation:) and String.init(stringInterpolationSegment:) directly yourself if you want (just try String(stringInterpolationSegment: 3.141) and String(stringInterpolation: "blah", "blah")), this doesn't really help you much. What you really need is a facade function that coordinates the calls to them. And unless there's a handy pre-existing function in the standard library that does exactly that which I've missed, I think you're out of luck. I suspect it's built into the compiler.
You could maybe write your own to achieve your goal, but a lot of effort since you'd have to break up the string you want to interpolate manually into bits and handle it yourself, calling the segment init in a loop. Also you'll hit problems with calling the combining function, since you can't splat an array into a variadic function call.
I don't think so. The compiler needs to be able to resolve the interpolated variable at compile time.
I'm not a Swift programmer, specifically, but I think you can workaround it to something pretty close to what you want using a Dictionary and standard string-replacing and splitting methods:
var replacement = [String: String]()
replacement["who"] = "programmers"
Having that, you can try to find the occurrences of "\(", reading what is next and prior to a ")", (this post can help with the split part, this one, with the replacing part), finding it in the dictionary, and reconstructing your string from the pieces you get.
this one works like a charm:
let who = "programmers"
let formatString = "Now is the time for all good %# to come to babble incoherently."
let aString = String(format: formatString, who)
Related
I am doing some processing of a string in Rust, and I need to be able to extract the last set of characters from that string. In other words, given a string like the following:
some|not|necessarily|long|name
I need to be able to get the last part of that string, namely "name" and put it into another String or a &str, in a manner like:
let last = call_some_function("some|not|necessarily|long|name");
so that last becomes equal to "name".
Is there a way to do this? Is there a string function that will allow this to be done easily? If not (after looking at the documentation, I doubt that there is), how would one do this in Rust?
While the answer from #effect is correct, it is not the most idiomatic nor the most performant way to do it. It'll walk the entire string and match all of the |s to reach the last. You can make it better, but there is a method of str that does exactly what you want - rsplit_once():
let (_, name) = s.rsplit_once('|').unwrap();
// Or
// let name = s.rsplit_once('|').unwrap().1;
//
// You can also use a multichar separator:
// let (_, name) = s.rsplit_once("|").unwrap();
// But in the case of a single character, a `char` type is likely to be more performant.
Playground.
You can use the String::split() method, which will return an iterator over the substrings split by that separator, and then use the Iterator::last() method to return the last element in the iterator, like so:
let s = String::from("some|not|necessarily|long|name");
let last = s.split('|').last().unwrap();
assert_eq!(last, "name");
Please also note that string slices (&str) also implement the split method, so you don't need to use std::String.
let s = "some|not|necessarily|long|name";
let last = s.split('|').last().unwrap();
assert_eq!(last, "name");
So the idea here is that I'm taking a .csv into a string and each value needs to be stored into a variable. I am unsure how to properly parse a string to do this.
My idea is a function that looks like
final char delim = ',';
int nextItem(String data, int startFrom) {
if (data.charAt(startFrom) != delim) {
return data.charAt(startFrom)
} else {
return nextItem(data, startFrom + 1);
}
}
so if I passed it something like
nextItem("45,621,9", 0);
it would return 45
and if I passed it
nextItem("45,621,9", 3);
it would return 621
I'm not sure if I have that setup properly to be recursive, but I could also use a For loop I suppose, only real stipulation is I can't use the Substring method.
Please don't use recursion for a matter that can be easily done iteratively. Recursion is expensive in terms of stack and calling frames: A very long string could produce a StackOverflowError.
I suggest you take a look to standard method indexOf of java.lang.String:
A good alternative is Regular Expressions.
You can seperate the words considering comma ',' as delimeter
Code
String[] nextItem(String data) {
String[] words=data.split(",");
return words;
}
This will return an array of strings that is the words in your input string. Then you can use the array in anyway you need.
Hope it helps ;)
Processing comes with a split() function that does exactly what you're describing.
From the reference:
String men = "Chernenko,Andropov,Brezhnev";
String[] list = split(men, ',');
// list[0] is now "Chernenko", list[1] is "Andropov"...
Behind the scenes it's using the String#split() function like H. Sodi's answer, but you should just use this function instead of defining your own.
I have an abstract class Writer which allows clients to write into something. Could be the screen, could be a file. Now, I try to create a derived class to write into a string.
I have two problems with the denoted line in method write(...):
It's probably very inefficient. Is there something like a string buffer in Matlab?
It writes escape sequences like \n plain into the string, instead of taking their actual meaning.
How can I get the denoted line properly?
Code:
classdef StringTextWriter < Writer
properties
str;
end
methods
function this = StringTextWriter()
% Init the write-target which is a string in our case.
% (Other Writer classes would maybe open a file.)
this.str = '';
end
function write(this, val)
% Write to writer target.
% (Other Writer classes would maybe use fprinf here for file write.)
% ?????????????????????????????
this.str = [this.str val]; % How to do this properly?
% ?????????????????????????????
end
end
end
To answer your questions point by point:
The closest notion to a string buffer would be a string cell. Instead of:
str = '';
str = [strbuf, 'abc\n'];
str = [strbuf, 'def\n'];
str = [strbuf, 'ghi\n'];
%// and so on...
one may use
strbuf = {};
strbuf{end+1} = 'abc\n';
strbuf{end+1} = 'def\n';
strbuf{end+1} = 'ghi\n';
%// and so on...
str = sprintf([strbuf{:}]); %// careful about percent signs and single quotes in string
the drawback being that you have to reconstruct the string every time you ask for it. This can be alleviated by setting a modified flag every time you add strings to the end of strbuf, resetting it every time you concatenate the strings, and memoizing the result of concatenation in the last line (rebuild if modified, or last result if not).
Further improvement can be achieved by choosing a better strategy for growing the strbuf cell array; probably this would be effective if you have a lot of write method calls.
The escape sequences are really linked to the <?>printf family and not to the string literals, so MATLAB in general doesn't care about them, but sprintf in particular might.
The plot
There is a rather complicatedly formatted string, like there's no such readable regex that parses it. And the aim is to get a specific substring for example, and to get it's original position. That substring is reached after parsing a bit, like trimming, removing the beginning something and searching the n-th element for example. I just want to demonstrate you the complexity with this example, otherwise it's pretty general.
For demonstration, see this rudimentary example. The way it is isn't really important, just to reach a pretty complicated parse model. Obviously, there can be more rule and you can write a simplier model as well.
FirstBlock{Index1, Index2} SecondBlock ThirdBlock
{ FirstBlock {Index1,Index2} SecondBlock}
{FirstBlock SecondBlock ThirdBlock FourthBlock}
I've tried to make as random as it could be. The parsing model is like:
string text = "{ FirstBlock {Index1,Index2} SecondBlock}";
text = text.Trim();
if (text.First() == '{')
{
text = text.SubString(1, text.Length - 2);
}
text = text.Trim();
string firstBlock = text.Split(new char[] { ' ', '{' })[0];
text = text.Remove(0, firstBlock.Length).Trim();
string indices = "";
if (text.First() == '{')
{
indices = text.Split(new char[] { '{', '}' })[0];
text = text.Remove(0, indices.Length).Trim();
}
string[] blocks = text.Split(' ');
The easy way
There is a way that is pretty easy to implement and straightforward. But does not give you the correct result sometimes. That way you parse the string and get the substring and then you make a re-search, for example string.IndexOf() and get the position. But if there are two match for example, you are given the first one even though it is not sure you wanted that one.
My notion
The way I think is quite elegant but still not consummate is to index the characters of the string at the beginning, then parse it, and eventually you end up with the proper characters and their position also. My problem there is that then you can't really use the functions the library gives, and I don't know a way to do that. Using the snippet above:
List<Tuple<int, char>> indexedText = text
.Select((ch, index) => new Tuple<int, char>(index, ch))
.ToList();
And with this structure you can still process the string without library methods but you are given the position indices eventually. For example, trim:
indexedText = indexedText
.SkipWhile(indexedChar => char.IsWhiteSpace(indexedChar.Item2))
.ToList();
The actual question
The question can either be a new solution or the way you can use library methods with indexed strings. The aim is to get the indices back after parsing a string. It is possible that there is a very simple way that is just out of my scope but I haven't found a proper solution yet. The solution I don't want is to simplify the parsing system, as I said it is just for demonstration.
This question already has answers here:
Is there an easy way to return a string repeated X number of times?
(21 answers)
Closed 9 years ago.
Just a curiosity I was investigating.
The matter: simply repeating (multiplying, someone would say) a string/character n times.
I know there is Enumerable.Repeat for this aim, but I was trying to do this without it.
LINQ in this case seems pretty useless, because in a query like
from X in "s" select X
the string "s" is being explored and so X is a char. The same is with extension methods, because for example "s".Aggregate(blablabla) would again work on just the character 's', not the string itself. For repeating the string something "external" would be needed, so I thought lambdas and delegates, but it can't be done without declaring a variable to assign the delegate/lambda expression to.
So something like defining a function and calling it inline:
( (a)=>{return " "+a;} )("a");
or
delegate(string a){return " "+a}(" ");
would give a "without name" error (and so no recursion, AFAIK, even by passing a possible lambda/delegate as a parameter), and in the end couldn't even be created by C# because of its limitations.
It could be that I'm watching this thing from the wrong perspective. Any ideas?
This is just an experiment, I don't care about performances, about memory use... Just that it is one line and sort of autonomous. Maybe one could do something with Copy/CopyTo, or casting it to some other collection, I don't know. Reflection is accepted too.
To repeat a character n-times you would not use Enumerable.Repeat but just this string constructor:
string str = new string('X', 10);
To repeat a string i don't know anything better than using string.Join and Enumerable.Repeat
string foo = "Foo";
string str = string.Join("", Enumerable.Repeat(foo, 10));
edit: you could use string.Concat instead if you need no separator:
string str = string.Concat( Enumerable.Repeat(foo, 10) );
If you're trying to repeat a string, rather than a character, a simple way would be to use the StringBuilder.Insert method, which takes an insertion index and a count for the number of repetitions to use:
var sb = new StringBuilder();
sb.Insert(0, "hi!", 5);
Console.WriteLine(sb.ToString());
Otherwise, to repeat a single character, use the string constructor as I've mentioned in the comments for the similar question here. For example:
string result = new String('-', 5); // -----
For the sake of completeness, it's worth noting that StringBuilder provides an overloaded Append method that can repeat a character, but has no such overload for strings (which is where the Insert method comes in). I would prefer the string constructor to the StringBuilder if that's all I was interested in doing. However, if I was already working with a StringBuilder, it might make sense to use the Append method to benefit from some chaining. Here's a contrived example to demonstrate:
var sb = new StringBuilder("This item is ");
sb.Insert(sb.Length, "very ", 2) // insert at the end to append
.Append('*', 3)
.Append("special")
.Append('*', 3);
Console.WriteLine(sb.ToString()); // This item is very very ***special***