fastcsv

CSV parser for golang that aims to be as fast as possible through the use of mmap and happens to provide a convenient interface.

Example usage:

var record struct {
    Name []byte `csv:"name"`
    Age  []byte `csv:"age"`
    Sex  []byte `csv:"sex"`
}

reader, err := NewFileReader("file.csv", ',', &record)
if err != nil {
    panic(err)
}
defer reader.Close()

for reader.Scan() {
    fmt.Printf("%s|%s|%s\n", record.Name, record.Age, record.Sex)
}

This setup currently gets 3.01 ns/row on my laptop, whereas the standard library gets 1872 ns/row. This can be tested with go test -bench=..

BUG: Currently the reader does not implemented the csv spec correctly yet; namely it does not handle quoted fields or escaped characters. It also always assumed rows are terminated by LF instead of CRLF.