Utility method for writing and reading from a multi-line string

[pdeffebach]

One thing that might be nice for Discourse posts and debugging in general is some sort of way to use CSV.jl to print to a multi-line string and then have CSV read such a string back into a data frame.

I was playing around with this just now and couldn't find the right combination of CSV.write(io), show(df, MIME("text/csv"), df) and take!(io) for this to work.

[quinnj]

It's pretty straightforward using IOBuffer, like:

io = IOBuffer()
CSV.write(io, df)
seekstart(io)
df2 = CSV.read(io, DataFrame)

[pdeffebach]

Sorry, I mean a way that a user can copy a multi-line string of their data into a discourse post, then I can read it into CSV from that.

That should be easy to do, right?

[quinnj]

Oh, yeah, probably something like:

julia> df = DataFrame(a=[1,2,3], b=[3.4, 5.6, 7.9], c=["hey","there","buddy"])
3×3 DataFrame
│ Row │ a     │ b       │ c      │
│     │ Int64 │ Float64 │ String │
├─────┼───────┼─────────┼────────┤
│ 1   │ 1     │ 3.4     │ hey    │
│ 2   │ 2     │ 5.6     │ there  │
│ 3   │ 3     │ 7.9     │ buddy  │

julia> io = IOBuffer()
IOBuffer(data=UInt8[...], readable=true, writable=true, seekable=true, append=false, size=0, maxsize=Inf, ptr=1, mark=-1)

julia> CSV.write(io, df)
IOBuffer(data=UInt8[...], readable=true, writable=true, seekable=true, append=false, size=40, maxsize=Inf, ptr=41, mark=-1)

julia> println(String(take!(io)))
a,b,c
1,3.4,hey
2,5.6,there
3,7.9,buddy

[pdeffebach]

But then, how do I read that into a data frame after it's printed?

[quinnj]

df = CSV.read(IOBuffer("""
a,b,c
1,3.4,hey
2,5.6,there
3,7.9,buddy
"""), DataFrame)

[pdeffebach]

Thanks! I will start recommending people do this on Discourse.