Feature request: substring_before_first(), substring_after_last(), etc.

[alt-graph]

It would be nice to have a family of substring extraction functions for situations like this:

auto extension = substring_after_last("readme.txt", ".");
REQUIRE(extension == "txt");

auto facility = substring_before_first("FACILITY/DEVICE/LOCATION/PROPERTY", "/");
REQUIRE(facility == "FACILITY");

The use cases for this functionality are frequent, and writing ad-hoc code for it is clumsy and makes the code hard to read.

To start off, I could imagine the following signatures:

std::string substring_after_first(std::string_view string, std::string_view separator);
std::string substring_after_last(std::string_view string, std::string_view separator);
std::string substring_before_first(std::string_view string, std::string_view separator);
std::string substring_before_last(std::string_view string, std::string_view separator);

We could add _sv versions returning string_view as well.

[Finii]

🤔

auto facility = split_sv<SmallVector<string_view>>("FACILITY/DEVICE/LOCATION/PROPERTY", "/")[0];

Does the separator need to be a string or can we limit to one character? Not that I know if that has any performance impact.

What I fear is that we will not get 4 but 8 functions, or course also a substring_after_first_sv() etc.
There must be a more generic approach.

Hmm, simulating x = '/'.join(sourcestring.split('/')[1..-1]) (for substring_after_first)

I have no good name, but maybe

ResultT extract_range(sourcestring, separator, first, last) where first and last are Python array slice indexes.

substring_after_first              extract_range(sourcestring, separator, 1, -1)
substring_after_last               extract_range(sourcestring, separator, -1, -1)
substring_before_first             extract_range(sourcestring, separator, 0, 0)
substring_before_last              extract_range(sourcestring, separator, -2, -1)

extract_doocs_location             extract_range(sourcestring, "/", 2, 2)