Add possibility to load file row by row

[Grufoony]

I am working with big csv files which I need to acquire on my own data structure.
The thing I am actually doing is to load the csv and then iterate through the number of rows and move data from the vectors returned as columns by this library but I find this particularly inefficient.
I was thinking about a function like iterrows which returns a map and you can read each cell by calling the column name with the correct type casting... would it be possible?

for (row : df.iterrows()) {
  auto data1 = row.get<type>("col1");
  auto data2 = row.get<type>("col2");
  ...
}

[d99kris]

Hi @Grufoony - this is not supported by rapidcsv currently, and there is no plans to support it. I would probably recommend checking out some other C++ CSV libraries, as many (most?) of them provide the type of API you are looking for. I haven't kept myself up-to-date on C++ CSV libraries, but you could try https://github.com/vincentlaucsb/csv-parser or consult Google or ChatGPT for a recommendation.