Closed lamorlette-cintoo closed 1 month ago
I used to construct CVSReader with a std::ifstream as follows:
CVSReader
std::ifstream
const std::filesystem::path file_path {"foo"}; std::ifstream input(file_path , std::ios::binary | std::ios::in); csv::CSVReader path_reader {input};
It is working perfectly fine with relatively small files, but it throws an exception with files bigger than 10 MB.
I found out that the same files work perfectly fine when using the other CVSReader constructor, taking the filename as parameter:
const std::filesystem::path file_path {"foo"}; csv::CSVReader path_reader {file_path.string()};
(Unfortunately I can't share with you an example because I don't own them.)
Duplicate of #180
Closing as duplicate of #180. While this is an issue, I am curious as to why people are simply not using the default memory-mapped CSVReader as recommended in the README.
CSVReader
I used to construct
CVSReader
with astd::ifstream
as follows:It is working perfectly fine with relatively small files, but it throws an exception with files bigger than 10 MB.
I found out that the same files work perfectly fine when using the other
CVSReader
constructor, taking the filename as parameter:(Unfortunately I can't share with you an example because I don't own them.)