A minimal implementation of CSV reader/writer for Julia. It reads and writes DataFrames from/to file on disk.

Installation: at the Julia REPL, using Pkg; Pkg.add(PackageSpec(url="https://github.com/bkamins/Nanocsv.jl"))

• always use " as quote character, "" represents a quote in quoted field
• you can change field separator using delim keyword argument
• when data is written to disk then string function is used to get its representation
• by default tries to parse columns to Int, Float64 using parse function and falls back to String; you can change list or sequence of tried parsers using parsers keyword argument
• handles missing values using na keyword argument; na is always unquoted when reading/writing (except if header line is present in a file as then it is read as is); if string equal to na is written it is always quoted; if value equal to na is quoted in file then it is treated as a valid value not missing

Writing CSV to disk:

write_csv(io::IO, df::DataFrame;
          delim::Char=',', na::AbstractString="")
write_csv(filename::AbstractString, df::DataFrame;
          delim::Char=',', na::AbstractString="")

Reading CSV from disk:

read_csv(io::IO;
         delim::Char=',', header::Bool=true, na::String="",
         parsers::Vector{DataType} = [Int, Float64],
         skiphead::Int=0, nrows::Union{Int, Nothing}=nothing)
read_csv(filename::AbstractString;
         delim::Char=',', header::Bool=true, na::String="",
         parsers::Vector{DataType} = [Int, Float64],
         skiphead::Int=0, nrows::Union{Int, Nothing}=nothing)