MapReduce word count using C++ transform_reduce() function with a parallel execution policy

I have a string of words and I would like to count the number of occurrences of each word, and store the results in a map. I want to use std::transform_reduce() in order to take advantage of it's parallel processing option and use it on a much larger dataset.

e.g.

std::string text = "apple orange banana apple apple orange";

std::istringstream iss(text);

// Put these words into a vector
// Define a vector (CTAD), use its range constructor and the std::istream_iterator as iterator
std::vector words(std::istream_iterator<std::string>(iss), {});

// Aim: Use transform_reduce() to populate an unordered_map that maps each word to its word count

std::unordered_map<std::string, int> wordCount;

// Count the occurrences using transform_reduce
std::transform_reduce(
    words.begin(),
    words.end(), 
    // use the unordered_map wordCount somehow?
    // [](){ lambda for reduce to populate the unordered_map wordCount}
    // [](){ lambda for transforming the words vector of data to pairs: e.g. ("apple", 1) }
);

How would I achieve this in C++17 or later using transform_reduce()?

Note: chatGPT had a stab at it, but its code didn't compile and I cannot see how to make it work.

I would use a std::ranges::subrange then use a range based for loop.

#include <unordered_map>
#include <string>
#include <sstream>
#include <iostream>
#include <ranges>
#include <iterator>



int main()
{
    using StreamWordIter = std::istream_iterator<std::string>;
    using SubRange       = std::ranges::subrange;

    std::string        text = "apple orange banana apple apple orange";    
    std::istringstream iss(text);

    std::unordered_map<std::string, int> wordCount;
    for (auto const& word: SubRange(StreamWordIter{iss}, StreamWordIter{})) {
        ++wordCount[word];
    }

    for (auto const& [word, count]: wordCount) {
        std::cout << "  Word: " << word << " => " << count << "\n";
    }
}