TL;DR I’m wondering if readr or tibble or anything base/tidyverse can make a tibble from a character vector of CSV rows (as if I had done read_lines() on some CSV file).

More explanation: I am collecting output (from stdout of a processx object) into a character vector, and this output is actually lines of a CSV, being created by the subprocess. I would like to make this into a tibble, but the only function I can find that does this is data.table::fread (via the ‘text’ arg). This is in a package and I don’t really want to take a dependency on data.table solely for this one function.

I’ve looked through the tidyverse docs and I can’t find anything. I realize I could write my own parser, and I started doing it with map_def and str_split but then I realized there are subtleties like “ignore the comma if it’s within quotes”, etc. maybe I’m just being lazy, but I don’t really want to implement a full csv parser when there are millions that already exist. Not to mention, speed concerns. This vector is sometimes 10k rows and I’m sure my homemade approach would not be the most optimized.

Any thoughts are welcome. Maybe I should just bite the bullet and either write the parser or use data.table but I feel like I can’t be the first person to want to do this and there must be something g built in that I’m missing. Thanks!

CodePudding user response：

Does read.csv not work?

read.csv(
  text = "a,b
          1,2",
  header = TRUE
)

#   a b
# 1 1 2