问题
I'm using the ben strasser C++ fast csv parser: https://github.com/ben-strasser/fast-cpp-csv-parser. It uses a variadic template to pass column values back to the while loop that processes the csv data:
io::CSVReader<2> in(csv_filename);
double x, y;
while(in.read_row(x,y)) {
//code with x and y
}
This calls the following function in the CSVReader class:
template<class ...ColType>
bool read_row(ColType& ...cols){
//snip
}
This works fine for me with my x and y values. However, I would like to expand this to use arbitrary dimensions. This means my data has a (known) number of columns I need to read. I would like to use something like this:
io::CSVReader<known_dimension> in(csvfname);
double data[known_dimension];
while(in.read_row(data)) {
//code with data[0],data[1],...,data[known_number]
}
However, this is not valid syntax. I need to "unpack" the array of doubles into separate arguments of pointers to my doubles. I'd like to do this without modifications to the fast csv parser.
回答1:
You can use std::integer_sequence for that purpose:
namespace Detail
{
template <typename Reader, typename T, std::size_t... I>
void read_row(Reader& in, T* data, std::index_sequence<I...>)
{
in.read_row(data[I]...); // A trick here
}
}
template <std::size_t N, typename T>
void read_row(io::CSVReader<N>& in, T* data)
{
Detail::read_row(in, data, std::make_index_sequence<N>{});
}
And of course, use like this:
int a[7];
io::CSVReader<7> r;
read_row(r, a);
"Working" example: link
For compiler "below" C++14 - integer_sequence
(actually just index_sequence
needed) is pretty easy to implement:
template <std::size_t... I>
class index_sequence {};
And make_index_sequence
not so easy - but also doable:
template <std::size_t N, std::size_t ...I>
struct make_index_sequence : make_index_sequence<N-1, N-1,I...> {};
template <std::size_t ...I>
struct make_index_sequence<0,I...> : index_sequence<I...> {};
来源:https://stackoverflow.com/questions/34929856/unpack-an-array-to-call-a-function-with-variadic-template