skipmissing(x) not defined

[Datseris]

I load a csv file with missing data using CSV, DataFrames; CSV.read(myfile) and the the missing data are represented using the word #NULL instead of missing, which is NOT the same as in the documentation page. (first red flag there)

Then, I try to use the function skipmissing, but it is undefined: UndefVarError: skipmissing not defined . I have also tried DataFrames.skipmissing, CSV.skipmissing.

I though that the stable documentation accompanies the version installed by Pkg.add and Pkg.update but this does not seem to be the case here?

In the meantime, can somebody tell me how I can drop NULL values from my data, because it is kind of super-crucial that I do it as fast as possible?

[nalimilan]

Can you post the versions of DataFrames, CSV, and Missings you are using? It looks like you don't have the latest version of Missings.

[Datseris]

I got - DataFrames 0.10.1....
Why isn't Pkg.update() updating it?

(edit, sorry if my tone is agressive, I am kind in tremendous pressure now since the documentation and what happens in my computer are two completely different things)

[nalimilan]

Because many dependencies are incompatible with it. See https://discourse.julialang.org/t/dataframes-0-11-released/7296/.

[Datseris]

I am looking at: #1232

If you can help me identify which packages I need to "remove" you are welcome...

40 required packages:
 - Atom                          0.6.5
 - BenchmarkTools                0.2.2
 - CSV                           0.1.5
 - Combinatorics                 0.5.0
 - DataFrames                    0.10.1
 - DataTables                    0.0.3
 - DiffEqBase                    2.6.1
 - Distributions                 0.15.0
 - Documenter                    0.12.3
 - ForwardDiff                   0.7.0
 - IJulia                        1.6.2
 - Interact                      0.6.3
 - Interpolations                0.7.2
 - IterTools                     0.1.0
 - JLD                           0.8.3
 - KernelDensity                 0.4.0
 - MIDI                          0.2.0+             master
 - Measurements                  0.5.0+             master
 - NLsolve                       0.13.0
 - NearestNeighbors              0.3.0
 - ODE                           0.7.0
 - OhMyREPL                      0.2.10
 - OrdinaryDiffEq                2.30.0
 - ParameterizedFunctions        2.3.0
 - PkgDev                        0.1.6
 - Plots                         0.13.1
 - ProfileView                   0.3.0
 - ProgressMeter                 0.5.2
 - PyCall                        1.15.0
 - PyPlot                        2.3.2
 - Reactive                      0.6.0
 - RecursiveArrayTools           0.12.4
 - RegionTrees                   0.1.0
 - Requires                      0.4.3
 - StaticArrays                  0.6.6
 - SymPy                         0.5.4
 - TaylorIntegration             0.2.0
 - TaylorSeries                  0.7.0
 - TypeSortedCollections         0.2.0
 - Unrolled                      0.0.1

On the other hand, I am looking at the documentation for version 0.10.1. But the problem is that the null values in my data right now are represented as :

149-element NullableArrays.NullableArray{Int64,1}:
 4    
 #NULL
 #NULL
 #NULL
 #NULL
 #NULL
 #NULL

a[2]
Nullable{Int64}()

which means that dropna doesn't do anything. Oh man this is really stressing :(

[nalimilan]

I'd avoid using CSV.jl with DataFrames 0.10.1. If you have a deadline, better keep using readtable to read CSV files, and work with DataArrays.

[Datseris]

Thank you thank you thank you with readtable dropna works!

[nalimilan]

Cool, closing then. Try moving to version 0.11 when you have more time.