UchidaMizuki / timbr

Forest Data Frames

timbr

timbr provides data frames for forest or tree data structures. You can create forest data structures from data frames and process them based on their hierarchies.

Installation

You can install the development version of timbr from GitHub with:

# the released version from CRAN:
install.packages("timbr")

# the development version from GitHub:
# install.packages("devtools")
devtools::install_github("UchidaMizuki/timbr")

Main Functions

The main functions provided by timbr are as follows,

• children()
• climb()
• leaves()
• traverse()
• rbind()

tidyverse methods

timbr provides some tidyverse methods as follows,

• mutate()
• summarise()
• select() and relocate()
• rows_update() and rows_patch()

Examples

library(timbr)
library(dplyr)

fr <- tidyr::expand_grid(key1 = letters[1:2],
                         key2 = letters[1:2],
                         key3 = letters[1:2]) |>  
  mutate(value = row_number()) |> 
  forest_by(key1, key2, key3)

fr_sum <- fr |> 
  summarise(value = sum(value)) |> 
  summarise(value = sum(value))

fr
#> # A forest: 8 nodes and 1 feature
#> # Groups:   key1, key2 [4]
#> # Trees:    
#> #   key3 [8]
#>   key1  key2  .        value
#>   <chr> <chr> <node>   <int>
#> 1 a     a     <key3> a     1
#> 2 a     a     <key3> b     2
#> 3 a     b     <key3> a     3
#> 4 a     b     <key3> b     4
#> 5 b     a     <key3> a     5
#> 6 b     a     <key3> b     6
#> 7 b     b     <key3> a     7
#> 8 b     b     <key3> b     8
fr_sum
#> # A forest: 14 nodes and 1 feature
#> # Trees:    
#> #   key1 [2]
#> #   └─key2 [4]
#> #     └─key3 [8]
#>   .        value
#>   <node>   <int>
#> 1 <key1> a    10
#> 2 <key1> b    26
children(fr_sum)
#> # A forest: 12 nodes and 1 feature
#> # Groups:   key1 [2]
#> # Trees:    
#> #   key2 [4]
#> #   └─key3 [8]
#>   key1  .        value
#>   <chr> <node>   <int>
#> 1 a     <key2> a     3
#> 2 a     <key2> b     7
#> 3 b     <key2> a    11
#> 4 b     <key2> b    15
fr_sum |> 
  climb(key3)
#> # A forest: 8 nodes and 1 feature
#> # Trees:    
#> #   key3 [8]
#>   .        value
#>   <node>   <int>
#> 1 <key3> a     1
#> 2 <key3> b     2
#> 3 <key3> a     3
#> 4 <key3> b     4
#> 5 <key3> a     5
#> 6 <key3> b     6
#> 7 <key3> a     7
#> 8 <key3> b     8

fr1 <- tidyr::expand_grid(key1 = letters[1:2], 
                          key2_1 = letters[1:2],
                          key3_1 = letters[1:2]) |> 
  mutate(value = row_number()) |> 
  forest_by(key1, key2_1, key3_1) |> 
  summarise(value = sum(value))

fr2 <- tidyr::expand_grid(key1 = letters[1:2], 
                          key2_2 = letters[1:2],
                          key3_2 = letters[1:2]) |> 
  mutate(value = row_number()) |> 
  forest_by(key1, key2_2, key3_2) |> 
  summarise(value = sum(value))

fr <- rbind(fr1, fr2)
fr_sum <- fr |> 
  summarise(value = sum(value))

fr
#> # A forest: 24 nodes and 1 feature
#> # Groups:   key1 [2]
#> # Trees:    
#> #   key2_1 [4]
#> #   └─key3_1 [8]
#> #   key2_2 [4]
#> #   └─key3_2 [8]
#>   key1  .          value
#>   <chr> <node>     <int>
#> 1 a     <key2_1> a     3
#> 2 a     <key2_1> b     7
#> 3 b     <key2_1> a    11
#> 4 b     <key2_1> b    15
#> 5 a     <key2_2> a     3
#> 6 a     <key2_2> b     7
#> 7 b     <key2_2> a    11
#> 8 b     <key2_2> b    15
fr_sum
#> # A forest: 26 nodes and 1 feature
#> # Trees:    
#> #   key1 [2]
#> #   ├─key2_1 [4]
#> #   │ └─key3_1 [8]
#> #   └─key2_2 [4]
#> #     └─key3_2 [8]
#>   .        value
#>   <node>   <int>
#> 1 <key1> a    20
#> 2 <key1> b    52
traverse(fr_sum,
         function(x, children) {
           x$value <- prod(children$value)
           x
         })
#> # A forest: 26 nodes and 1 feature
#> # Trees:    
#> #   key1 [2]
#> #   ├─key2_1 [4]
#> #   │ └─key3_1 [8]
#> #   └─key2_2 [4]
#> #     └─key3_2 [8]
#>   .          value
#>   <node>     <int>
#> 1 <key1> a     576
#> 2 <key1> b 2822400