[julia-users] Re: Writing a subset DataFrame to file is 220 times slower than saving the whole DataFrame

[Fred]

Wow ! Impressive difference, it is 100 time faster without sub() :


s = df[df[:rank_PV].<=r_max,:]
@time write_results(s, name, "significant", sep, h)
Saving... significant/Stat.csv
 0.704880 seconds (7.14 M allocations: 164.687 MB, 2.40% gc time)

s = sub(df, (df[:rank_PV] .<= r_max))
@time write_results(s, name, "significant", sep, h)
Saving...    significant/Stat.csv
 86.797541 seconds (21.98 M allocations: 104.215 GB, 4.13% gc time)



Thanks Alex !