Julia のデータフレームで Query

直にデータフレームを操作するより楽なのかな？

using Query
using RDatasets
mtcars = dataset("datasets", "mtcars");
names(mtcars)

12-element Array{String,1}:
 "Model"
 "MPG"
 "Cyl"
 "Disp"
 "HP"
 "DRat"
 "WT"
 "QSec"
 "VS"
 "AM"
 "Gear"
 "Carb"

@collect で，条件に合ったものをデータフレームとして返す。

q1 = @from i  in mtcars begin
    @where i.AM == 1
    @select {i.Model, i.Cyl, i.MPG, i.WT}
    @collect DataFrame
end

13 rows × 4 columns

	Model	Cyl	MPG	WT
	String	Int64	Float64	Float64
1	Mazda RX4	6	21.0	2.62
2	Mazda RX4 Wag	6	21.0	2.875
3	Datsun 710	4	22.8	2.32
4	Fiat 128	4	32.4	2.2
5	Honda Civic	4	30.4	1.615
6	Toyota Corolla	4	33.9	1.835
7	Fiat X1-9	4	27.3	1.935
8	Porsche 914-2	4	26.0	2.14
9	Lotus Europa	4	30.4	1.513
10	Ford Pantera L	8	15.8	3.17
11	Ferrari Dino	6	19.7	2.77
12	Maserati Bora	8	15.0	3.57
13	Volvo 142E	4	21.4	2.78

 

普通のデータフレームになる。

using Statistics
cor(q1[!, 3], q1[!, 4])

-0.9089147883714981

cor(Matrix(q1[:, 2:4]))

3×3 Array{Float64,2}:
  1.0       -0.825998   0.847024
 -0.825998   1.0       -0.908915
  0.847024  -0.908915   1.0

@collect しなければ，イテレータを返す。

q2 = @from i  in mtcars begin
    @where i.AM == 1
    @select {i.Model, i.WT}
end

Model	WT
"Mazda RX4"	2.62
"Mazda RX4 Wag"	2.875
"Datsun 710"	2.32
"Fiat 128"	2.2
"Honda Civic"	1.615
"Toyota Corolla"	1.835
"Fiat X1-9"	1.935
"Porsche 914-2"	2.14
"Lotus Europa"	1.513
"Ford Pantera L"	3.17
⋮	⋮

... with more rows.

イテレータは for や 内包表記で使える。

for i in q2
    println("$(i.Model): weight is $(i.WT)")
end

Mazda RX4: weight is 2.62
Mazda RX4 Wag: weight is 2.875
Datsun 710: weight is 2.32
Fiat 128: weight is 2.2
Honda Civic: weight is 1.615
Toyota Corolla: weight is 1.835
Fiat X1-9: weight is 1.935
Porsche 914-2: weight is 2.14
Lotus Europa: weight is 1.513
Ford Pantera L: weight is 3.17
Ferrari Dino: weight is 2.77
Maserati Bora: weight is 3.57
Volvo 142E: weight is 2.78

[i.Model for i in q2 if i.WT > 2.5]

6-element Array{String,1}:
 "Mazda RX4"
 "Mazda RX4 Wag"
 "Ford Pantera L"
 "Ferrari Dino"
 "Maserati Bora"
 "Volvo 142E"