C++ Weekly - Ep 340 - Finally! A Simple String Split in C++!

☟☟ Awesome T-Shirts! Sponsors! Books! ☟☟

FlexiSpot Pro Plus Standing Desk E7
#MyFlexiDesk #flexiblewfh

T-SHIRTS AVAILABLE!

WANT MORE JASON?

SUPPORT THE CHANNEL

GET INVOLVED

JASON'S BOOKS

► C++23 Best Practices

► C++ Best Practices

JASON'S PUZZLE BOOKS

► Object Lifetime Puzzlers Book 1

► Object Lifetime Puzzlers Book 2

► Object Lifetime Puzzlers Book 3

► Copy and Reference Puzzlers Book 1

► Copy and Reference Puzzlers Book 2

► Copy and Reference Puzzlers Book 3

► OpCode Puzzlers Book 1


RECOMMENDED BOOKS

AWESOME PROJECTS

O'Reilly VIDEOS

[fukawitribe]

There's some very nice aspects of C++, and it's been my daily driver for a lot of years - but the fact that it's 2022 and splitting a string _still_ looks like the dog's dinner, and even after all that you have to jump through hoops to use the chunks you get back does say something about it... ;)

[dmdjt]

I have the urge to applaud you for the respectful way you did the ad!

[N....]

I usually split using string_view anyway, nice to see ranges has such an elegant solution!

[ABaumstumpf]

I still can not understand why the C++ members seem so opposed to a simple stringsplit/concatenate method.

[anon_y_mousse]

This is great. You getting sponsorships will hopefully lead to more videos, I hope.

[prahTV]

Most languages' definition of simple:

print "hello world".split(" ")

C++ definition of simple:

constexpr static auto str = std::string_view{"hello world"} | std::ranges::views::split(' ');
for (const auto &string : auto{split_strings}) {
std::cout << std::string_view{string};
}

[mocloun]

It is still far from easy! This would be easy: auto splitted = std::splitt(string, " ");
And a "range of ranges" is not a standard container. The output should be std::vector<std::string> or similar.

[lucasrangit]

tell me more use cases for constexpr string views please

[nites]

That auto{} in the for loop that "makes the copy explicit" is completely new for me. What's the name of that feature?

[chigozie]

I remember also adding my answer to another question asking about string splitting in c++, on stackoverflow.
My answer was lazy in nature because it created an iterator over the tokens rather than forcing one to specify a container to collect the tokens into, and the tokens are generated as the iterator is consumed.

I was as surprised then (about 8 years ago) as I am now, that something this simple is still an issue for C++.
I can't wait until Carbon becomes mature, I bet they already have a neat solution in mind.

[Xilefian]

Would ‘constinit’ make sense for that ‘split_strings’?

[yato]

what if words are separated by multiple spaces? How does it handle it? Does it filter out empty ranges or you have to do it yourself?

[v-for-victory]

Splitting strings is still stress in C++ but at least this a good approach to solve it.

[franciscogerardohernandezr]

I find the Matlab style quite friendly. To all experts, why porting a similar library with such simplicity has not been done? Any caveats I'm missing?

[Key_Capz_]

This might be useful for better way for parsing a whole text document and contain it for indexing and mapping.

[ohwow]

Finally! Lol. But can we also split based on other delimiters and not just space or tab characters?

[nothke]

How does a "range of ranges" actually work? Since it can't know how many views there will be, it must allocate dynamic memory, right? It's essentially like a vector of ranges?

[ChaotikmindSrc]

i personally very rarely use std::string, using my own class, which has a split method XD

[progammler]

The 'splits' I get from this are string_views and not strings. They are invalidated as soon as the input string is modified/destroyed (without warnings!). I still need to keep track of the data myself. If you want to keep the splits as members in a class then you loose the copy-/move-ctors unless you keep the data outside the object or you manually make new strings from the string_views. I believe the use of string_views to be quite limited, it would be nice if they were more than just pointers into a string (or less = just offsets not tied to a string instead of pointers). I'd rather be able to store string_views inside the string so that they stay copyable/moveable.

[TheSulross]

still looking for the video where will be chiming in with take on Google's Carbon language (such as it is right now)