Transformation to DataFrames

Transformation to DataFrames#

Split-apply-combine

using DataFrames

Grouping a data frame#

groupby

x = DataFrame(id=[1, 2, 3, 4, 1, 2, 3, 4], id2=[1, 2, 1, 2, 1, 2, 1, 2], v=rand(8))

8×3 DataFrame

Row	id	id2	v
	Int64	Int64	Float64
1	1	1	0.274951
2	2	2	0.00502915
3	3	1	0.27654
4	4	2	0.497871
5	1	1	0.913103
6	2	2	0.994002
7	3	1	0.124519
8	4	2	0.123377

groupby(x, :id)

GroupedDataFrame with 4 groups based on key: id

First Group (2 rows): id = 1

Row	id	id2	v
	Int64	Int64	Float64
1	1	1	0.274951
2	1	1	0.913103

⋮

Last Group (2 rows): id = 4

Row	id	id2	v
	Int64	Int64	Float64
1	4	2	0.497871
2	4	2	0.123377

groupby(x, [])

GroupedDataFrame with 1 group based on key:

First Group (8 rows):

Row	id	id2	v
	Int64	Int64	Float64
1	1	1	0.274951
2	2	2	0.00502915
3	3	1	0.27654
4	4	2	0.497871
5	1	1	0.913103
6	2	2	0.994002
7	3	1	0.124519
8	4	2	0.123377

gx2 = groupby(x, [:id, :id2])

GroupedDataFrame with 4 groups based on keys: id, id2

First Group (2 rows): id = 1, id2 = 1

Row	id	id2	v
	Int64	Int64	Float64
1	1	1	0.274951
2	1	1	0.913103

⋮

Last Group (2 rows): id = 4, id2 = 2

Row	id	id2	v
	Int64	Int64	Float64
1	4	2	0.497871
2	4	2	0.123377

get the parent DataFrame

parent(gx2)

8×3 DataFrame

Row	id	id2	v
	Int64	Int64	Float64
1	1	1	0.274951
2	2	2	0.00502915
3	3	1	0.27654
4	4	2	0.497871
5	1	1	0.913103
6	2	2	0.994002
7	3	1	0.124519
8	4	2	0.123377

back to the DataFrame, but in a different order of rows than the original

vcat(gx2...)

8×3 DataFrame

Row	id	id2	v
	Int64	Int64	Float64
1	1	1	0.274951
2	1	1	0.913103
3	2	2	0.00502915
4	2	2	0.994002
5	3	1	0.27654
6	3	1	0.124519
7	4	2	0.497871
8	4	2	0.123377

the same as above

DataFrame(gx2)

8×3 DataFrame

Row	id	id2	v
	Int64	Int64	Float64
1	1	1	0.274951
2	1	1	0.913103
3	2	2	0.00502915
4	2	2	0.994002
5	3	1	0.27654
6	3	1	0.124519
7	4	2	0.497871
8	4	2	0.123377

drop grouping columns when creating a data frame

DataFrame(gx2, keepkeys=false)

8×1 DataFrame

Row	v
	Float64
1	0.274951
2	0.913103
3	0.00502915
4	0.994002
5	0.27654
6	0.124519
7	0.497871
8	0.123377

vector of names of grouping variables

groupcols(gx2)

2-element Vector{Symbol}:
 :id
 :id2

and non-grouping variables

valuecols(gx2)

1-element Vector{Symbol}:
 :v

group indices in parent(gx2)

groupindices(gx2)

8-element Vector{Union{Missing, Int64}}:
 1
 2
 3
 4
 1
 2
 3
 4

kgx2 = keys(gx2)

4-element DataFrames.GroupKeys{DataFrames.GroupedDataFrame{DataFrames.DataFrame}}:
 GroupKey: (id = 1, id2 = 1)
 GroupKey: (id = 2, id2 = 2)
 GroupKey: (id = 3, id2 = 1)
 GroupKey: (id = 4, id2 = 2)

You can index into a GroupedDataFrame like to a vector or to a dictionary. The second form accepts GroupKey, NamedTuple or a Tuple.

gx2

GroupedDataFrame with 4 groups based on keys: id, id2

First Group (2 rows): id = 1, id2 = 1

Row	id	id2	v
	Int64	Int64	Float64
1	1	1	0.274951
2	1	1	0.913103

⋮

Last Group (2 rows): id = 4, id2 = 2

Row	id	id2	v
	Int64	Int64	Float64
1	4	2	0.497871
2	4	2	0.123377

k = keys(gx2)[1]

GroupKey: (id = 1, id2 = 1)

ntk = NamedTuple(k)

(id = 1, id2 = 1)

tk = Tuple(k)

(1, 1)

the operations below produce the same result and are proformant

gx2[1], gx2[k], gx2[ntk], gx2[tk]

(2×3 SubDataFrame
 Row │ id     id2    v        
     │ Int64  Int64  Float64  
─────┼────────────────────────
   1 │     1      1  0.274951
   2 │     1      1  0.913103, 2×3 SubDataFrame
 Row │ id     id2    v        
     │ Int64  Int64  Float64  
─────┼────────────────────────
   1 │     1      1  0.274951
   2 │     1      1  0.913103, 2×3 SubDataFrame
 Row │ id     id2    v        
     │ Int64  Int64  Float64  
─────┼────────────────────────
   1 │     1      1  0.274951
   2 │     1      1  0.913103, 2×3 SubDataFrame
 Row │ id     id2    v        
     │ Int64  Int64  Float64  
─────┼────────────────────────
   1 │     1      1  0.274951
   2 │     1      1  0.913103)

handling missing values

x = DataFrame(id=[missing, 5, 1, 3, missing], x=1:5)

5×2 DataFrame

Row	id	x
	Int64?	Int64
1	missing	1
2	5	2
3	1	3
4	3	4
5	missing	5

by default groups include missing values and their order is not guaranteed

groupby(x, :id)

GroupedDataFrame with 4 groups based on key: id

First Group (1 row): id = 1

Row	id	x
	Int64?	Int64
1	1	3

⋮

Last Group (2 rows): id = missing

Row	id	x
	Int64?	Int64
1	missing	1
2	missing	5

but we can change it; now they are sorted

groupby(x, :id, sort=true, skipmissing=true)

GroupedDataFrame with 3 groups based on key: id

First Group (1 row): id = 1

Row	id	x
	Int64?	Int64
1	1	3

⋮

Last Group (1 row): id = 5

Row	id	x
	Int64?	Int64
1	5	2

and now they are in the order they appear in the source data frame

groupby(x, :id, sort=false)

GroupedDataFrame with 4 groups based on key: id

First Group (2 rows): id = missing

Row	id	x
	Int64?	Int64
1	missing	1
2	missing	5

⋮

Last Group (1 row): id = 3

Row	id	x
	Int64?	Int64
1	3	4

Performing transformations#

by group using combine, select, select!, transform, and transform!

using Statistics
using Chain

x = DataFrame(id=rand('a':'d', 100), v=rand(100))

100×2 DataFrame

75 rows omitted

Row	id	v
	Char	Float64
1	c	0.86718
2	d	0.00553355
3	a	0.792594
4	d	0.145016
5	a	0.723366
6	c	0.373478
7	a	0.156386
8	a	0.774747
9	b	0.319325
10	b	0.577269
11	a	0.179759
12	b	0.399449
13	c	0.0531153
⋮	⋮	⋮
89	d	0.314694
90	b	0.694085
91	a	0.148528
92	d	0.90592
93	d	0.364386
94	d	0.307045
95	c	0.803253
96	b	0.941827
97	d	0.4352
98	c	0.874392
99	c	0.0448102
100	b	0.814797

apply a function to each group of a data frame combine keeps as many rows as are returned from the function

@chain x begin
    groupby(:id)
    combine(:v => mean)
end

4×2 DataFrame

Row	id	v_mean
	Char	Float64
1	c	0.541363
2	d	0.449724
3	a	0.520584
4	b	0.552102

x.id2 = axes(x, 1)

Base.OneTo(100)

Select and transform keep as many rows as are in the source data frame and in correct order. Additionally, transform keeps all columns from the source.

@chain x begin
    groupby(:id)
    transform(:v => mean)
end

100×4 DataFrame

75 rows omitted

Row	id	v	id2	v_mean
	Char	Float64	Int64	Float64
1	c	0.86718	1	0.541363
2	d	0.00553355	2	0.449724
3	a	0.792594	3	0.520584
4	d	0.145016	4	0.449724
5	a	0.723366	5	0.520584
6	c	0.373478	6	0.541363
7	a	0.156386	7	0.520584
8	a	0.774747	8	0.520584
9	b	0.319325	9	0.552102
10	b	0.577269	10	0.552102
11	a	0.179759	11	0.520584
12	b	0.399449	12	0.552102
13	c	0.0531153	13	0.541363
⋮	⋮	⋮	⋮	⋮
89	d	0.314694	89	0.449724
90	b	0.694085	90	0.552102
91	a	0.148528	91	0.520584
92	d	0.90592	92	0.449724
93	d	0.364386	93	0.449724
94	d	0.307045	94	0.449724
95	c	0.803253	95	0.541363
96	b	0.941827	96	0.552102
97	d	0.4352	97	0.449724
98	c	0.874392	98	0.541363
99	c	0.0448102	99	0.541363
100	b	0.814797	100	0.552102

note that combine reorders rows by group of GroupedDataFrame

@chain x begin
    groupby(:id)
    combine(:id2, :v => mean)
end

100×3 DataFrame

75 rows omitted

Row	id	id2	v_mean
	Char	Int64	Float64
1	c	1	0.541363
2	c	6	0.541363
3	c	13	0.541363
4	c	21	0.541363
5	c	22	0.541363
6	c	25	0.541363
7	c	31	0.541363
8	c	33	0.541363
9	c	36	0.541363
10	c	50	0.541363
11	c	51	0.541363
12	c	54	0.541363
13	c	55	0.541363
⋮	⋮	⋮	⋮
89	b	68	0.552102
90	b	69	0.552102
91	b	70	0.552102
92	b	72	0.552102
93	b	76	0.552102
94	b	77	0.552102
95	b	80	0.552102
96	b	85	0.552102
97	b	87	0.552102
98	b	90	0.552102
99	b	96	0.552102
100	b	100	0.552102

we give a custom name for the result column

@chain x begin
    groupby(:id)
    combine(:v => mean => :res)
end

4×2 DataFrame

Row	id	res
	Char	Float64
1	c	0.541363
2	d	0.449724
3	a	0.520584
4	b	0.552102

you can have multiple operations

@chain x begin
    groupby(:id)
    combine(:v => mean => :res1, :v => sum => :res2, nrow => :n)
end

4×4 DataFrame

Row	id	res1	res2	n
	Char	Float64	Float64	Int64
1	c	0.541363	13.5341	25
2	d	0.449724	10.7934	24
3	a	0.520584	13.5352	26
4	b	0.552102	13.8026	25

Additional notes:

select! and transform! perform operations in-place
The general syntax for transformation is source_columns => function => target_column
if you pass multiple columns to a function they are treated as positional arguments
ByRow and AsTable work exactly like discussed for operations on data frames in 05_columns.ipynb
you can automatically groupby again the result of combine, select etc. by passing ungroup=false keyword argument to them
similarly keepkeys keyword argument allows you to drop grouping columns from the resulting data frame

It is also allowed to pass a function to all these functions (also - as a special case, as a first argument). In this case the return value can be a table. In particular it allows for an easy dropping of groups if you return an empty table from the function.

If you pass a function you can use a do block syntax. In case of passing a function it gets a SubDataFrame as its argument.

Here is an example:

combine(groupby(x, :id)) do sdf
    n = nrow(sdf)
    n < 25 ? DataFrame() : DataFrame(n=n) ## drop groups with low number of rows
end

3×2 DataFrame

Row	id	n
	Char	Int64
1	c	25
2	a	26
3	b	25

You can also produce multiple columns in a single operation:

df = DataFrame(id=[1, 1, 2, 2], val=[1, 2, 3, 4])

4×2 DataFrame

Row	id	val
	Int64	Int64
1	1	1
2	1	2
3	2	3
4	2	4

@chain df begin
    groupby(:id)
    combine(:val => (x -> [x]) => AsTable)
end

2×3 DataFrame

Row	id	x1	x2
	Int64	Int64	Int64
1	1	1	2
2	2	3	4

@chain df begin
    groupby(:id)
    combine(:val => (x -> [x]) => [:c1, :c2])
end

2×3 DataFrame

Row	id	c1	c2
	Int64	Int64	Int64
1	1	1	2
2	2	3	4

It is easy to unnest the column into multiple columns,

df = DataFrame(a=[(p=1, q=2), (p=3, q=4)])
select(df, :a => AsTable)

2×2 DataFrame

Row	p	q
	Int64	Int64
1	1	2
2	3	4

automatic column names generated

df = DataFrame(a=[[1, 2], [3, 4]])
select(df, :a => AsTable)

2×2 DataFrame

Row	x1	x2
	Int64	Int64
1	1	2
2	3	4

custom column names generated

select(df, :a => [:C1, :C2])

2×2 DataFrame

Row	C1	C2
	Int64	Int64
1	1	2
2	3	4

Finally, observe that one can conveniently apply multiple transformations using broadcasting:

df = DataFrame(id=repeat(1:10, 10), x1=1:100, x2=101:200)

100×3 DataFrame

75 rows omitted

Row	id	x1	x2
	Int64	Int64	Int64
1	1	1	101
2	2	2	102
3	3	3	103
4	4	4	104
5	5	5	105
6	6	6	106
7	7	7	107
8	8	8	108
9	9	9	109
10	10	10	110
11	1	11	111
12	2	12	112
13	3	13	113
⋮	⋮	⋮	⋮
89	9	89	189
90	10	90	190
91	1	91	191
92	2	92	192
93	3	93	193
94	4	94	194
95	5	95	195
96	6	96	196
97	7	97	197
98	8	98	198
99	9	99	199
100	10	100	200

@chain df begin
    groupby(:id)
    combine([:x1, :x2] .=> minimum)
end

10×3 DataFrame

Row	id	x1_minimum	x2_minimum
	Int64	Int64	Int64
1	1	1	101
2	2	2	102
3	3	3	103
4	4	4	104
5	5	5	105
6	6	6	106
7	7	7	107
8	8	8	108
9	9	9	109
10	10	10	110

@chain df begin
    groupby(:id)
    combine([:x1, :x2] .=> [minimum maximum])
end

10×5 DataFrame

Row	id	x1_minimum	x2_minimum	x1_maximum	x2_maximum
	Int64	Int64	Int64	Int64	Int64
1	1	1	101	91	191
2	2	2	102	92	192
3	3	3	103	93	193
4	4	4	104	94	194
5	5	5	105	95	195
6	6	6	106	96	196
7	7	7	107	97	197
8	8	8	108	98	198
9	9	9	109	99	199
10	10	10	110	100	200

Aggregation of a data frame using mapcols#

x = DataFrame(rand(10, 10), :auto)

10×10 DataFrame

Row	x1	x2	x3	x4	x5	x6	x7	x8	x9	x10
	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64
1	0.969356	0.885177	0.761078	0.496073	0.581544	0.0793956	0.531403	0.864059	0.607208	0.249705
2	0.844299	0.691145	0.780836	0.628189	0.320369	0.642588	0.978763	0.814136	0.465593	0.220491
3	0.0102794	0.444881	0.973632	0.574662	0.664609	0.0182701	0.173161	0.651498	0.861454	0.363842
4	0.896474	0.692801	0.0800644	0.255091	0.399623	0.312342	0.223707	0.63687	0.754104	0.747419
5	0.081352	0.427961	0.0648244	0.87397	0.230091	0.839031	0.106634	0.648813	0.478749	0.609716
6	0.286477	0.355591	0.402054	0.0750566	0.435071	0.425402	0.607839	0.458655	0.323706	0.0805402
7	0.282261	0.637702	0.639553	0.979943	0.399984	0.838833	0.580271	0.839842	0.0535565	0.0232844
8	0.817089	0.0949748	0.0509811	0.115348	0.877574	0.0681337	0.148046	0.29752	0.958121	0.601349
9	0.414355	0.731444	0.725687	0.106148	0.346254	0.717215	0.835049	0.970882	0.261405	0.913889
10	0.978132	0.773139	0.469198	0.160293	0.176057	0.581785	0.289374	0.987972	0.903049	0.788043

mapcols(mean, x)

1×10 DataFrame

Row	x1	x2	x3	x4	x5	x6	x7	x8	x9	x10
	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64
1	0.558007	0.573482	0.494791	0.426478	0.443118	0.452299	0.447425	0.717025	0.566695	0.459828

Mapping rows and columns using eachcol and eachrow#

map a function over each column and return a vector

map(mean, eachcol(x))

10-element Vector{Float64}:
5580074671136023
5734816749457339
49479074789167593
42647752642362546
44311770773949916
45229947965039824
447424705841956
7170247908651912
5666947144849279
4598278711412883

an iteration returns a Pair with column name and values

foreach(c -> println(c[1], ": ", mean(c[2])), pairs(eachcol(x)))

x1: 0.5580074671136023
x2: 0.5734816749457339
x3: 0.49479074789167593
x4: 0.42647752642362546
x5: 0.44311770773949916
x6: 0.45229947965039824
x7: 0.447424705841956
x8: 0.7170247908651912
x9: 0.5666947144849279
x10: 0.4598278711412883

now the returned value is DataFrameRow which works as a NamedTuple but is a view to a parent DataFrame

map(r -> r.x1 / r.x2, eachrow(x))

10-element Vector{Float64}:
0950983038719277
2215946823759614
023105980733285687
2939839832212443
19009186137341755
8056369667857306
44262260834169254
603218574055887
5664889266820277
2651444081680663

it prints like a data frame, only the caption is different so that you know the type of the object

er = eachrow(x)
er.x1 ## you can access columns of a parent data frame directly

10-element Vector{Float64}:
9693560197606523
8442993309808335
010279419744060547
8964735498352752
08135197669469263
28647738846072346
28226121332184273
8170886247220703
4143549589180511
9781321886978206

it prints like a data frame, only the caption is different so that you know the type of the object

ec = eachcol(x)

10×10 DataFrameColumns

Row	x1	x2	x3	x4	x5	x6	x7	x8	x9	x10
	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64	Float64
1	0.969356	0.885177	0.761078	0.496073	0.581544	0.0793956	0.531403	0.864059	0.607208	0.249705
2	0.844299	0.691145	0.780836	0.628189	0.320369	0.642588	0.978763	0.814136	0.465593	0.220491
3	0.0102794	0.444881	0.973632	0.574662	0.664609	0.0182701	0.173161	0.651498	0.861454	0.363842
4	0.896474	0.692801	0.0800644	0.255091	0.399623	0.312342	0.223707	0.63687	0.754104	0.747419
5	0.081352	0.427961	0.0648244	0.87397	0.230091	0.839031	0.106634	0.648813	0.478749	0.609716
6	0.286477	0.355591	0.402054	0.0750566	0.435071	0.425402	0.607839	0.458655	0.323706	0.0805402
7	0.282261	0.637702	0.639553	0.979943	0.399984	0.838833	0.580271	0.839842	0.0535565	0.0232844
8	0.817089	0.0949748	0.0509811	0.115348	0.877574	0.0681337	0.148046	0.29752	0.958121	0.601349
9	0.414355	0.731444	0.725687	0.106148	0.346254	0.717215	0.835049	0.970882	0.261405	0.913889
10	0.978132	0.773139	0.469198	0.160293	0.176057	0.581785	0.289374	0.987972	0.903049	0.788043

you can access columns of a parent data frame directly

ec.x1

10-element Vector{Float64}:
9693560197606523
8442993309808335
010279419744060547
8964735498352752
08135197669469263
28647738846072346
28226121332184273
8170886247220703
4143549589180511
9781321886978206

Transposing#

you can transpose a data frame using permutedims:

df = DataFrame(reshape(1:12, 3, 4), :auto)

3×4 DataFrame

Row	x1	x2	x3	x4
	Int64	Int64	Int64	Int64
1	1	4	7	10
2	2	5	8	11
3	3	6	9	12

df.names = ["a", "b", "c"]

3-element Vector{String}:
 "a"
 "b"
 "c"

permutedims(df, :names)

4×4 DataFrame

Row	names	a	b	c
	String	Int64	Int64	Int64
1	x1	1	2	3
2	x2	4	5	6
3	x3	7	8	9
4	x4	10	11	12

This notebook was generated using Literate.jl.

Transformation to DataFrames

Contents

Transformation to DataFrames#

Grouping a data frame#

Performing transformations#

Aggregation of a data frame using mapcols#

Mapping rows and columns using eachcol and eachrow#

Transposing#