Rapidcsv VS Fast C++ CSV Parser

https://github.com/nlohmann/json and https://github.com/d99kris/rapidcsv are single header file include libs

babashka includes a huge array of libraries for CSV, JSON, HTTP, and way more. I doubt jank would include those, not for technical reasons, but that it's out-of-scope. If you wanted CSV support, for example, you'll probably need to link/compile-in https://github.com/d99kris/rapidcsv or something

Have you searched on the internet? A quick search gives https://github.com/d99kris/rapidcsv, amongst others.

Sorry that I still have a life. Had to analyse and mangle many large csv files and switched from Python and its csv module to: https://github.com/ben-strasser/fast-cpp-csv-parser and C++ containers/types. Python is just ineffable slow in comparison (as I wrote from >1hr to <1s).

where csv_reader is a CSVReader type from: https://github.com/ben-strasser/fast-cpp-csv-parser, we're reading iris.data, a CSV file which contains double,double,double,double,string values.

Good night everybody, I was trying to work with some csv file that had some hebrew characters, therefore I need to use std::wstring, but the ones I found didn't supported wide characters, like this one.\ Does anybody here knows any csv parser that works with wide chars? Thanks in advance.

*Go with my first instinct and do it in C++ anyways? I already use python at work and I'm wondering if a tool featuring a C++ parser as the back-end and something like pysimplegui as the front end could be a game-changer at my job.

This one comes to mind: https://github.com/ben-strasser/fast-cpp-csv-parser

I did consider the parser returning std::optional, but I think it is situational whether one would look nicer than the other. For example, using value_or() in the first example would not work since I do not want to print anything if the row is invalid, so I would still have to check the optional, and than add another line to decompose the tuple resulting in more lines. Notice that p.valid() can also be used to check if the file was open within the constructor. Fetching error messages would also be slightly more complicated. As for exceptions, I do no think it could work at all in this iteration loop since it would break the loop even if I catch it, and again, file not open would need to be handled too, and making it throw would require me to enclose the whole parser in a try/catch block since the constructor would throw that exception, one of the problems I had with the fast-cpp-csv-parser. Its all a trade-off, but I think I will stick with p.valid(), tho I will consider it still.

fast-cpp-csv-parser - 2.022 +- 0.025 [s]