std.csv Performance Review

[Jon Degenhardt via Digitalmars-d]

On Saturday, 3 June 2017 at 04:25:27 UTC, Jesse Phillips wrote:
> Author here:
>
> The discussion[1] and articles[2] around "Faster Command Line 
> Tools" had me trying out std.csv for the task.
>
> [snip]
>
> Keep in mind that the file itself has 10,512,769 rows of data 
> with four columns. Now I've talked to std.csv's performance in 
> the past, probably with the author of the fast command line 
> tools.
> 
> [snip]

I'm the author of the TSV tools. I'd be happy to provide insights 
I've gained from this exercise to help improve std.csv. I did 
examine std.csv when I first starting writing the TSV tools, but 
decided they weren't appropriate for what I was trying to do. I 
don't remember the specifics at this point though. CSV imposes 
many requirements on an implementation that TSV does not, which 
makes it challenging to produce an implementation optimal for 
both.

--Jon