Create a tree programmatically

Let’s start by creating a tree programmatically. We do this by creating Node objects, and linking them together so as to define the parent-child relationships.

In this example, we are looking at a company, Acme Inc., and the tree reflects its organisational structure. The root (level 1) is the company. On level 2, the nodes represent departments, and the leaves of the tree represent projects that the company is considering for next year:

library(data.tree)

acme <- Node$new("Acme Inc.")
  accounting <- acme$AddChild("Accounting")
    software <- accounting$AddChild("New Software")
    standards <- accounting$AddChild("New Accounting Standards")
  research <- acme$AddChild("Research")
    newProductLine <- research$AddChild("New Product Line")
    newLabs <- research$AddChild("New Labs")
  it <- acme$AddChild("IT")
    outsource <- it$AddChild("Outsource")
    agile <- it$AddChild("Go agile")
    goToR <- it$AddChild("Switch to R")

print(acme)

##                           levelName
## 1  Acme Inc.                       
## 2   ¦--Accounting                  
## 3   ¦   ¦--New Software            
## 4   ¦   °--New Accounting Standards
## 5   ¦--Research                    
## 6   ¦   ¦--New Product Line        
## 7   ¦   °--New Labs                
## 8   °--IT                          
## 9       ¦--Outsource               
## 10      ¦--Go agile                
## 11      °--Switch to R

As you can see from the previous example, each Node is identified by its name, i.e. the argument you pass into the Node$new(name) constructor. The name needs to be unique among siblings, such that paths to Nodes are unambiguous.

Node inherits from R6 reference class. This has the following implications:

1. You can call methods on a Node in OO style, e.g. acme$Get("name")
2. Node exhibits reference semantics. Thus, multiple variables in R can point to the same Node, and modifying a Node will modify it for all referencing variables. In the above code example, both acme$IT and it reference the same object. This is different from the value semantics, which is much more widely used in R.

Create a tree from a data.frame

Creating a tree programmatically is useful especially in the context of algorithms. However, most times you will create a tree by conversion. This could be by conversion from a nested list-of-lists, by conversion from another R tree-structure (e.g. an ape phylo), or by conversion from a data.frame. For more details on all the options, type ?as.Node and refer to the See Also section.

One of the most common conversions is the one from a data.frame in table format. The following code illustrates this. We load the GNI2014 data from the treemap package. This data.frame is in table format, meaning that each row will represent a leaf in the data.tree structure:

library(treemap)
data(GNI2014)
head(GNI2014)

##   iso3          country     continent population    GNI
## 3  BMU          Bermuda North America      67837 106140
## 4  NOR           Norway        Europe    4676305 103630
## 5  QAT            Qatar          Asia     833285  92200
## 6  CHE      Switzerland        Europe    7604467  88120
## 7  MAC Macao SAR, China          Asia     559846  76270
## 8  LUX       Luxembourg        Europe     491775  75990

Let’s convert that into a data.tree structure! We start by defining a pathString. The pathString describes the hierarchy by defining a path from the root to each leaf. In this example, the hierarchy comes very naturally:

GNI2014$pathString <- paste("world", 
                            GNI2014$continent, 
                            GNI2014$country, 
                            sep = "/")

Once our pathString is defined, conversion to Node is very easy:

population <- as.Node(GNI2014)
print(population, "iso3", "population", "GNI", limit = 20)

##                                 levelName iso3 population    GNI
## 1  world                                               NA     NA
## 2   ¦--North America                                   NA     NA
## 3   ¦   ¦--Bermuda                         BMU      67837 106140
## 4   ¦   ¦--United States                   USA  313973000  55200
## 5   ¦   ¦--Canada                          CAN   33487208  51630
## 6   ¦   ¦--Bahamas, The                    BHS     309156  20980
## 7   ¦   ¦--Trinidad and Tobago             TTO    1310000  20070
## 8   ¦   ¦--Puerto Rico                     PRI    3971020  19310
## 9   ¦   ¦--Barbados                        BRB     284589  15310
## 10  ¦   ¦--St. Kitts and Nevis             KNA      40131  14920
## 11  ¦   ¦--Antigua and Barbuda             ATG      85632  13300
## 12  ¦   ¦--Panama                          PAN    3360474  11130
## 13  ¦   ¦--Costa Rica                      CRI    4253877  10120
## 14  ¦   ¦--Mexico                          MEX  111211789   9870
## 15  ¦   ¦--Grenada                         GRD      90739   7910
## 16  ¦   ¦--St. Lucia                       LCA     160267   7260
## 17  ¦   ¦--Dominica                        DMA      72660   6930
## 18  ¦   ¦--St. Vincent and the Grenadines  VCT     104574   6610
## 19  ¦   ¦--Dominican Republic              DOM    9650054   6040
## 20  ¦   °--... 7 nodes w/ 0 sub                        NA     NA
## 21  °--... 6 nodes w/ 171 sub                          NA     NA

This is a simple example, and more options are available. Type ?FromDataFrameTable for all the details.

Create a tree from a file

Often, trees are created from one of many file formats. When developing this package, We opted for a multi-step approach, meaning that you first import the file into one of the well-known R data structures. Then you convert these into a data.tree structure. For example, typical import patterns could be:

• csv -> data.frame in table format (?read.csv) -> data.tree (?as.Node.data.frame)
• Newick -> ape phylo (?ape::read.tree) -> data.tree (?as.Node.phylo )
• csv -> data.frame in network format (?read.csv) -> data.tree (c.f. ?FromDataFrameNetwork)
• yaml -> list of lists (?yaml::yaml.load) -> data.tree (?as.Node.list)
• json -> list of lists (e.g. ?jsonlite::fromJSON) -> data.tree (?as.Node.list)

If you have a choice, we recommend you consider yaml format to store and share your hierarchies. It is concise, human-readable, and very easy to convert to a data.tree. An example is provided here for illustration. The data represents what platforms and OS versions a group of students use:

library(yaml)
yaml <- "
name: OS Students 2014/15
OS X:
  Yosemite:
    users: 16
  Leopard:
    users: 43
Linux:
  Debian:
    users: 27
  Ubuntu:
    users: 36
Windows:
  W7:
    users: 31
  W8:
    users: 32
  W10:
    users: 4
"

osList <- yaml.load(yaml)
osNode <- as.Node(osList)
print(osNode, "users")

##              levelName users
## 1  OS Students 2014/15    NA
## 2   ¦--OS X               NA
## 3   ¦   ¦--Yosemite       16
## 4   ¦   °--Leopard        43
## 5   ¦--Linux              NA
## 6   ¦   ¦--Debian         27
## 7   ¦   °--Ubuntu         36
## 8   °--Windows            NA
## 9       ¦--W7             31
## 10      ¦--W8             32
## 11      °--W10             4

In cases where your leaf elements have no attributes, you might want to interpret them as nodes, and not as attributes. In such cases, you can use interpretNullAsList = TRUE to convert these into Nodes (instead of attributes).

For example:

library(yaml)
yaml <- "
name: OS Students 2014/15
OS X:
  Yosemite:
  Leopard:
Linux:
  Debian:
  Ubuntu:
Windows:
  W7:
  W8:
  W10:
"

osList <- yaml.load(yaml)
osNode <- as.Node(osList, interpretNullAsList = TRUE)
osNode$printFormatters <- list(h = "\u2500" , v = "\u2502", l = "\u2514", j = "\u251C")
print(osNode, "users")

##              levelName users
## 1  OS Students 2014/15    NA
## 2  ├─OS X                 NA
## 3  │├─Yosemite            NA
## 4  │└─Leopard             NA
## 5  ├─Linux                NA
## 6  │├─Debian              NA
## 7  │└─Ubuntu              NA
## 8  └─Windows              NA
## 9  ├─W7                   NA
## 10 ├─W8                   NA
## 11 └─W10                  NA

Actives Examples (aka Properties)

Actives look and feel like attributes, but they are dynamically evaluated. They are documented in the Node documentation, which is accessed by typing ?Node.

Remember our population example:

print(population, limit = 15)

##                        levelName
## 1  world                        
## 2   ¦--North America            
## 3   ¦   ¦--Bermuda              
## 4   ¦   ¦--United States        
## 5   ¦   ¦--Canada               
## 6   ¦   ¦--Bahamas, The         
## 7   ¦   ¦--Trinidad and Tobago  
## 8   ¦   ¦--Puerto Rico          
## 9   ¦   ¦--Barbados             
## 10  ¦   ¦--St. Kitts and Nevis  
## 11  ¦   ¦--Antigua and Barbuda  
## 12  ¦   ¦--Panama               
## 13  ¦   ¦--Costa Rica           
## 14  ¦   ¦--Mexico               
## 15  ¦   °--... 12 nodes w/ 0 sub
## 16  °--... 6 nodes w/ 176 sub

population$isRoot

## [1] TRUE

population$height

## [1] 3

population$count

## [1] 7

population$totalCount

## [1] 196

population$attributes

## character(0)

population$attributesAll

## [1] "GNI"        "continent"  "country"    "iso3"       "population"

population$averageBranchingFactor

## [1] 24.375

The naming convention of the package is that attributes and actives are lower case, whereas methods are upper / CamelCase. RStudio and other IDEs work well with data.tree. If you have a Node, simply type myNode$ + SPACE to get a list of available attributes, actives and methods.

OO-Style Methods Examples

Examples of OO-Style methods

You will find more information on these examples below.

Get will traverse the tree and collect specific values for the Nodes it traverses:

sum(population$Get("population", filterFun = isLeaf))

## [1] 6683146875

Prune traverses the tree and keeps only the subtrees for which the pruneFun returns TRUE.

Prune(population, pruneFun = function(x) !x$isLeaf || x$population > 1000000)

## [1] 39

Note that the Prune function has side-effects, as it acts on the original population object. The population sum is now smaller:

sum(population$Get("population", filterFun = isLeaf), na.rm = TRUE)

## [1] 6669737814

Traditional R Methods

popClone <- Clone(acme)

Traditional S3 generics are available especially for conversion:

as.data.frame(acme)

##                           levelName
## 1  Acme Inc.                       
## 2   ¦--Accounting                  
## 3   ¦   ¦--New Software            
## 4   ¦   °--New Accounting Standards
## 5   ¦--Research                    
## 6   ¦   ¦--New Product Line        
## 7   ¦   °--New Labs                
## 8   °--IT                          
## 9       ¦--Outsource               
## 10      ¦--Go agile                
## 11      °--Switch to R

Though there is also a more specialised non-generic version:

ToDataFrameNetwork(acme)

##          from                       to
## 1   Acme Inc.               Accounting
## 2   Acme Inc.                 Research
## 3   Acme Inc.                       IT
## 4  Accounting             New Software
## 5  Accounting New Accounting Standards
## 6    Research         New Product Line
## 7    Research                 New Labs
## 8          IT                Outsource
## 9          IT                 Go agile
## 10         IT              Switch to R

Navigation by path

The most natural form of climbing a tree is to climb by path:

acme$IT$Outsource

##   levelName
## 1 Outsource

acme$Research$`New Labs`

##   levelName
## 1  New Labs

Navigation by position

However, there is a number of other ways to get to a specific Node. We can access the children of a Node directly through Node$children:

acme$children[[1]]$children[[2]]$name

## [1] "New Accounting Standards"

Navigation by attributes

Furthermore, we can not only navigate by name, but also by other attributes. This is achieved with the Climb method. The name of each ... argument designates the field, and the value matches against Nodes. Each argument refers to the subsequent level to climb. In this example, Climb takes acme’s child at position 1 (i.e. Accounting), then it takes Accounting's child called New Software:

acme$Climb(position = 1, name = "New Software")$path

## [1] "Acme Inc."    "Accounting"   "New Software"

As a shortcut, you can climb multiple levels with a single argument:

tree <- CreateRegularTree(5, 5)
tree$Climb(position = c(2, 3, 4))$path

## [1] "1"       "1.2"     "1.2.3"   "1.2.3.4"

Finally, you can even combine. The following example starts on the root, then looks for child at position 2, then for its child at position 3. Next, we move to the child having name = “1.2.3.4”, and finally its child having name “1.2.3.4.5”:

tree$Climb(position = c(2, 3), name = c("1.2.3.4", "1.2.3.4.5"))$path

## [1] "1"         "1.2"       "1.2.3"     "1.2.3.4"   "1.2.3.4.5"

Custom attributes in constructor

An alternative, often convenient way to assign custom attributes is in the constructor, or in the Node$AddChild method:

birds <- Node$new("Aves", vulgo = "Bird")
birds$AddChild("Neognathae", vulgo = "New Jaws", species = 10000)
birds$AddChild("Palaeognathae", vulgo = "Old Jaws", species = 60)
print(birds, "vulgo", "species")

##           levelName    vulgo species
## 1 Aves                  Bird      NA
## 2  ¦--Neognathae    New Jaws   10000
## 3  °--Palaeognathae Old Jaws      60

Custom attributes as function

Nothing stops you from setting a function as a field. This calculates a value dynamically, i.e. whenever a field is accessed in tree traversal. For example, you can add a new Node to your structure, and the function will reflect this. Think of this as a hierarchical spreadsheet, in which you can set formulas into cells.

Consider the following example:

birds$species <- function(self) sum(sapply(self$children, function(x) x$species))
print(birds, "species")

##           levelName species
## 1 Aves                10060
## 2  ¦--Neognathae      10000
## 3  °--Palaeognathae      60

data.tree maps the self argument to the Node at hand. Thus, you must name the argument self.

Now, let’s assume we discover a new species. Then, the species on the root adjusts dynamically:

birds$Palaeognathae$species <- 61
print(birds, "species")

##           levelName species
## 1 Aves                10061
## 2  ¦--Neognathae      10000
## 3  °--Palaeognathae      61

This, together with the Set method and recursion, becomes a very powerful tool, as we’ll see later.

Basic Printing

Basic printing is easy, as you surely have noted in the previous sections. print displays a tree in a tree-grid view. On the left, you have the hierarchy. Then you have a column per variable you want to print:

print(acme, "cost", "p")

##                           levelName    cost    p
## 1  Acme Inc.                             NA   NA
## 2   ¦--Accounting                        NA   NA
## 3   ¦   ¦--New Software             1000000 0.50
## 4   ¦   °--New Accounting Standards  500000 0.75
## 5   ¦--Research                          NA   NA
## 6   ¦   ¦--New Product Line         2000000 0.25
## 7   ¦   °--New Labs                  750000 0.90
## 8   °--IT                                NA   NA
## 9       ¦--Outsource                 400000 0.20
## 10      ¦--Go agile                  250000 0.05
## 11      °--Switch to R                50000 1.00

For more advanced printing, you have a few options.

Formatters

You can use formatters to output a variable in a certain way. You can use formatters in two ways:

• You can set them on a Node using the SetFormat method. If you do this, then the formatter will be picked up as a default formatter whenever you print, Get, convert to data.frame, etc. Formatters can be set on any Node in a data.tree structure act on any descendant. So you can overwrite a formatter for a sub-tree.
• You can add an explicit ad-hoc formatter to the Get method (see below). This will overwrite default formatters previously set via the SetFormat method. You can also set the formatter to identity to void a default formatter.

Setting a formatter using the SetFormat method:

SetFormat(acme, "p", formatFun = FormatPercent)
SetFormat(acme, "cost", formatFun = function(x) FormatFixedDecimal(x, digits = 2))
print(acme, "cost", "p")

##                           levelName       cost        p
## 1  Acme Inc.                                           
## 2   ¦--Accounting                                      
## 3   ¦   ¦--New Software             1000000.00  50.00 %
## 4   ¦   °--New Accounting Standards  500000.00  75.00 %
## 5   ¦--Research                                        
## 6   ¦   ¦--New Product Line         2000000.00  25.00 %
## 7   ¦   °--New Labs                  750000.00  90.00 %
## 8   °--IT                                              
## 9       ¦--Outsource                 400000.00  20.00 %
## 10      ¦--Go agile                  250000.00   5.00 %
## 11      °--Switch to R                50000.00 100.00 %

Printing using Get

Formatting with the Get method overwrites any formatters found along the path:

data.frame(cost = acme$Get("cost", format = function(x) FormatFixedDecimal(x, 2)),
           p = acme$Get("p", format = FormatPercent))

##                                cost        p
## Acme Inc.                                   
## Accounting                                  
## New Software             1000000.00  50.00 %
## New Accounting Standards  500000.00  75.00 %
## Research                                    
## New Product Line         2000000.00  25.00 %
## New Labs                  750000.00  90.00 %
## IT                                          
## Outsource                 400000.00  20.00 %
## Go agile                  250000.00   5.00 %
## Switch to R                50000.00 100.00 %

plot

data.tree is mainly a data structure. As it is easy to convert data.tree structures to other formats, you have access to a large number of tools to plot a data.tree structure. For example, you can plot a data.tree structure as a dendrogram, as an ape tree, as a treeview, etc. Additionally, data.tree also provides its own plotting facility. It is built on GraphViz/DiagrammeR, and you can access these features via the plot and ToGraphViz functions. Note that DiagrammeR is not required to use data.tree, so plot only works if DiagrammeR is installed on your system. For example:

plot(acme)

acme

Styling

Similar to formatters for printing, you can style your tree and store the styling directly in the tree, for later use:

SetGraphStyle(acme, rankdir = "TB")
SetEdgeStyle(acme, arrowhead = "vee", color = "grey35", penwidth = 2)
SetNodeStyle(acme, style = "filled,rounded", shape = "box", fillcolor = "GreenYellow", 
            fontname = "helvetica", tooltip = GetDefaultTooltip)
SetNodeStyle(acme$IT, fillcolor = "LightBlue", penwidth = "5px")
plot(acme)

acme

For details on the styling attributes, see http://graphviz.org/Documentation.php .

Note that, by default, most Node style attributes will be inherited. Though, for example, label will not be inherited. However, inheritance can be avoided for all style attributes, as for the Accounting node in the following example:

SetNodeStyle(acme$Accounting, inherit = FALSE, fillcolor = "Thistle", 
             fontcolor = "Firebrick", tooltip = "This is the accounting department")
plot(acme)

acme

Use Do to set style on specific nodes:

Do(acme$leaves, function(node) SetNodeStyle(node, shape = "egg"))
plot(acme)

acme

Other Visualisations

However, there are also endless other possibilities to visualise data.tree structures. There are more examples in the applications vignette. Type vignette('applications', package = "data.tree").

plot(as.dendrogram(CreateRandomTree(nodes = 20)), center = TRUE)

library(igraph)
plot(as.igraph(acme, directed = TRUE, direction = "climb"))

library(networkD3)
acmeNetwork <- ToDataFrameNetwork(acme, "name")
simpleNetwork(acmeNetwork[-3], fontSize = 12)

fileName <- system.file("extdata", "useR15.csv", package="data.tree")
useRdf <- read.csv(fileName, stringsAsFactors = FALSE)
#define the hierarchy (Session/Room/Speaker)
useRdf$pathString <- paste("useR", useRdf$session, useRdf$room, useRdf$speaker, sep="|")
#convert to Node
useRtree <- as.Node(useRdf, pathDelimiter = "|")

#plot with networkD3
useRtreeList <- ToListExplicit(useRtree, unname = TRUE)
radialNetwork( useRtreeList)

Converting to data.frame

As you saw just above, creating a data.frame is easy.

Again, note that we always call such methods on the root Node of a data.tree structure, or on the root Node of a subtree:

acmedf <- as.data.frame(acme)
as.data.frame(acme$IT)

##         levelName
## 1 IT             
## 2  ¦--Outsource  
## 3  ¦--Go agile   
## 4  °--Switch to R

The same can be achieved by using the more specialised method:

ToDataFrameTree(acme)

We can also add field values of the Nodes as columns to the data.frame:

ToDataFrameTree(acme, "level", "cost")

##                           levelName level    cost
## 1  Acme Inc.                            1      NA
## 2   ¦--Accounting                       2      NA
## 3   ¦   ¦--New Software                 3 1000000
## 4   ¦   °--New Accounting Standards     3  500000
## 5   ¦--Research                         2      NA
## 6   ¦   ¦--New Product Line             3 2000000
## 7   ¦   °--New Labs                     3  750000
## 8   °--IT                               2      NA
## 9       ¦--Outsource                    3  400000
## 10      ¦--Go agile                     3  250000
## 11      °--Switch to R                  3   50000

Note that it is not required that the field is set on each and every Node.

Other data frame conversions are:

ToDataFrameTable(acme, "pathString", "cost")

##                                      pathString    cost
## 1             Acme Inc./Accounting/New Software 1000000
## 2 Acme Inc./Accounting/New Accounting Standards  500000
## 3           Acme Inc./Research/New Product Line 2000000
## 4                   Acme Inc./Research/New Labs  750000
## 5                        Acme Inc./IT/Outsource  400000
## 6                         Acme Inc./IT/Go agile  250000
## 7                      Acme Inc./IT/Switch to R   50000

ToDataFrameNetwork(acme, "cost")

##          from                       to    cost
## 1   Acme Inc.               Accounting      NA
## 2   Acme Inc.                 Research      NA
## 3   Acme Inc.                       IT      NA
## 4  Accounting             New Software 1000000
## 5  Accounting New Accounting Standards  500000
## 6    Research         New Product Line 2000000
## 7    Research                 New Labs  750000
## 8          IT                Outsource  400000
## 9          IT                 Go agile  250000
## 10         IT              Switch to R   50000

And, finally, we can also put attributes of our nodes in a column, based on a type discriminator. This sounds more complicated then what it is. Consider the default discriminator, level:

ToDataFrameTypeCol(acme, 'cost')

##     level_1    level_2                  level_3    cost
## 1 Acme Inc. Accounting             New Software 1000000
## 2 Acme Inc. Accounting New Accounting Standards  500000
## 3 Acme Inc.   Research         New Product Line 2000000
## 4 Acme Inc.   Research                 New Labs  750000
## 5 Acme Inc.         IT                Outsource  400000
## 6 Acme Inc.         IT                 Go agile  250000
## 7 Acme Inc.         IT              Switch to R   50000

Let’s look at a somewhat more advanced example. First, let’s assume that for the outsourcing project, we have two separate possibilities: Outsourcing to India or outsourcing to Poland:

acme$IT$Outsource$AddChild("India")
acme$IT$Outsource$AddChild("Poland")

Now, with this slightly more complex tree structure, the level is not a usefully discriminator anymore, because some projects are in level 3, while the new projects are in level 4. For this reason, we introduce a type field on our node objects: A node type can be a company (root only), a department (Accounting, Research, and IT), a program (Oursource), and a project (the rest, i.e. all the leaves):

acme$Set(type = c('company', 'department', 'project', 'project', 'department', 'project', 'project', 'department', 'program', 'project', 'project', 'project', 'project'))

Our tree now looks like this:

print(acme, 'type')

##                           levelName       type
## 1  Acme Inc.                           company
## 2   ¦--Accounting                   department
## 3   ¦   ¦--New Software                project
## 4   ¦   °--New Accounting Standards    project
## 5   ¦--Research                     department
## 6   ¦   ¦--New Product Line            project
## 7   ¦   °--New Labs                    project
## 8   °--IT                           department
## 9       ¦--Outsource                   program
## 10      ¦   ¦--India                   project
## 11      ¦   °--Poland                  project
## 12      ¦--Go agile                    project
## 13      °--Switch to R                 project

We can now create a data.frame in which we have one column per distinct type value. Namely, a company column, a department column, a program column, and a project column. Note that the columns are not hardcoded, but derived dynamically from your data in the tree structure:

ToDataFrameTypeCol(acme, type = 'type', prefix = NULL)

##     company department   program                  project
## 1 Acme Inc. Accounting      <NA>             New Software
## 2 Acme Inc. Accounting      <NA> New Accounting Standards
## 3 Acme Inc.   Research      <NA>         New Product Line
## 4 Acme Inc.   Research      <NA>                 New Labs
## 5 Acme Inc.         IT Outsource                    India
## 6 Acme Inc.         IT Outsource                   Poland
## 7 Acme Inc.         IT      <NA>                 Go agile
## 8 Acme Inc.         IT      <NA>              Switch to R

Converting to List of Lists

List of lists are useful for various use cases:

• as an intermediate step in converting to JSON, XML, YAML
• for functions that take a lol as an input. This is especially the case for visualisations and charts, e.g with many html widgets
• to save a data.tree structure as an R object (see performance considerations below)

data(acme)
str(as.list(acme$IT))

## List of 3
##  $ Outsource  :List of 2
##   ..$ cost: num 4e+05
##   ..$ p   : num 0.2
##  $ Go agile   :List of 2
##   ..$ cost: num 250000
##   ..$ p   : num 0.05
##  $ Switch to R:List of 2
##   ..$ cost: num 50000
##   ..$ p   : num 1

str(ToListExplicit(acme$IT, unname = FALSE, nameName = "id", childrenName = "dependencies"))

## List of 2
##  $ id          : chr "IT"
##  $ dependencies:List of 3
##   ..$ Outsource  :List of 3
##   .. ..$ id  : chr "Outsource"
##   .. ..$ cost: num 4e+05
##   .. ..$ p   : num 0.2
##   ..$ Go agile   :List of 3
##   .. ..$ id  : chr "Go agile"
##   .. ..$ cost: num 250000
##   .. ..$ p   : num 0.05
##   ..$ Switch to R:List of 3
##   .. ..$ id  : chr "Switch to R"
##   .. ..$ cost: num 50000
##   .. ..$ p   : num 1

Converting to other objects

There are also conversions to igraph objects, to phylo / ape, to dendrogram, and others. For details, see ?as.phylo.Node, ?as.dendrogram.Node, ?as.igraph.Node.

Traversal order

The Get method can traverse the tree in various ways. This is called traversal order.

Pre-Order

The default traversal mode is pre-order.

pre-order

This is what is used e.g. in print:

print(acme, "level")

##                           levelName level
## 1  Acme Inc.                            1
## 2   ¦--Accounting                       2
## 3   ¦   ¦--New Software                 3
## 4   ¦   °--New Accounting Standards     3
## 5   ¦--Research                         2
## 6   ¦   ¦--New Product Line             3
## 7   ¦   °--New Labs                     3
## 8   °--IT                               2
## 9       ¦--Outsource                    3
## 10      ¦--Go agile                     3
## 11      °--Switch to R                  3

Post-Order

The post-order traversal mode returns children first, returning parents only after all its children have been traversed and returned:

post-order

We can use it like this on the Get method:

acme$Get('level', traversal = "post-order")

##             New Software New Accounting Standards               Accounting 
##                        3                        3                        2 
##         New Product Line                 New Labs                 Research 
##                        3                        3                        2 
##                Outsource                 Go agile              Switch to R 
##                        3                        3                        3 
##                       IT                Acme Inc. 
##                        2                        1

This is useful if your parent’s value depends on the children, as we’ll see below.

data.frame(level = agile$Get('level', traversal = "ancestor"))

##           level
## Go agile      3
## IT            2
## Acme Inc.     1

Filter and Prune

You can add a filter and/or a prune function to the Get method. These functions have to take a Node as an input, and return TRUE if the Node should be considered, and FALSE otherwise. The difference between the pruneFun and the filterFun is that filters act only on specific nodes, whereas if the pruneFun returns FALSE, then the entire sub-tree spanned by the Node is ignored.

For example:

acme$Get('name', pruneFun = function(x) x$position <= 2)

##                  Acme Inc.                 Accounting 
##                "Acme Inc."               "Accounting" 
##               New Software   New Accounting Standards 
##             "New Software" "New Accounting Standards" 
##                   Research           New Product Line 
##                 "Research"         "New Product Line" 
##                   New Labs 
##                 "New Labs"

There are also some convenient filter functions available in the package, such as isLeaf, isRoot, isNotLeaf, etc.

acme$Get('name', filterFun = isLeaf)

##               New Software   New Accounting Standards 
##             "New Software" "New Accounting Standards" 
##           New Product Line                   New Labs 
##         "New Product Line"                 "New Labs" 
##                  Outsource                   Go agile 
##                "Outsource"                 "Go agile" 
##                Switch to R 
##              "Switch to R"

Attributes

The attribute parameter determines what is collected. This is called attribute, but it should not be confused with R’s concept of object attributes (e.g. ?attributes). In this context, an attribute can be either:

• the name of a Node field
• the name of a Node method or active
• a function, whose first argument must be a Node

Throughout this document, we refer to attribute in this sense.

acme$Get('name')

##                  Acme Inc.                 Accounting 
##                "Acme Inc."               "Accounting" 
##               New Software   New Accounting Standards 
##             "New Software" "New Accounting Standards" 
##                   Research           New Product Line 
##                 "Research"         "New Product Line" 
##                   New Labs                         IT 
##                 "New Labs"                       "IT" 
##                  Outsource                   Go agile 
##                "Outsource"                 "Go agile" 
##                Switch to R 
##              "Switch to R"

ExpectedCost <- function(node, adjustmentFactor = 1) {
  return ( node$cost * node$p * adjustmentFactor)
}

acme$Get(ExpectedCost, adjustmentFactor = 0.9, filterFun = isLeaf)

##             New Software New Accounting Standards         New Product Line 
##                   450000                   337500                   450000 
##                 New Labs                Outsource                 Go agile 
##                   607500                    72000                    11250 
##              Switch to R 
##                    45000

Cost <- function(node) {
  result <- node$cost
  if(length(result) == 0) result <- sum(sapply(node$children, Cost))
  return (result)
}

print(acme, "p", cost = Cost)

##                           levelName    p    cost
## 1  Acme Inc.                          NA 4950000
## 2   ¦--Accounting                     NA 1500000
## 3   ¦   ¦--New Software             0.50 1000000
## 4   ¦   °--New Accounting Standards 0.75  500000
## 5   ¦--Research                       NA 2750000
## 6   ¦   ¦--New Product Line         0.25 2000000
## 7   ¦   °--New Labs                 0.90  750000
## 8   °--IT                             NA  700000
## 9       ¦--Outsource                0.20  400000
## 10      ¦--Go agile                 0.05  250000
## 11      °--Switch to R              1.00   50000

Assigning values

acme$Set(id = 1:acme$totalCount)

print(acme, "id")

##                           levelName id
## 1  Acme Inc.                         1
## 2   ¦--Accounting                    2
## 3   ¦   ¦--New Software              3
## 4   ¦   °--New Accounting Standards  4
## 5   ¦--Research                      5
## 6   ¦   ¦--New Product Line          6
## 7   ¦   °--New Labs                  7
## 8   °--IT                            8
## 9       ¦--Outsource                 9
## 10      ¦--Go agile                 10
## 11      °--Switch to R              11

The Set method can take multiple vectors as an input, and, optionally, you can define the name of the attribute. Finally, just as for the Get method, the traversal order is important for the Set.

secretaries <- c(3, 2, 8)
employees <- c(52, 43, 51)
acme$Set(secretaries, 
         emps = employees,
         filterFun = function(x) x$level == 2)
print(acme, "emps", "secretaries", "id")

##                           levelName emps secretaries id
## 1  Acme Inc.                          NA          NA  1
## 2   ¦--Accounting                     52           3  2
## 3   ¦   ¦--New Software               NA          NA  3
## 4   ¦   °--New Accounting Standards   NA          NA  4
## 5   ¦--Research                       43           2  5
## 6   ¦   ¦--New Product Line           NA          NA  6
## 7   ¦   °--New Labs                   NA          NA  7
## 8   °--IT                             51           8  8
## 9       ¦--Outsource                  NA          NA  9
## 10      ¦--Go agile                   NA          NA 10
## 11      °--Switch to R                NA          NA 11

Deleting attributes

The Set method can also be used to assign a single value directly to all Nodes traversed. For example, to remove the avgExpectedCost, we assign NULL on each node, using the fact that the Set recycles:

acme$Set(avgExpectedCost = NULL)

However, note that setting a field to NULL will not make it gone for good. You will still see it:

acme$attributesAll

## [1] "avgExpectedCost" "cost"            "id"              "emps"           
## [5] "secretaries"     "p"

In order remove it completely, you can use the RemoveAttribute method:

acme$Do(function(node) node$RemoveAttribute("avgExpectedCost"))

Using Set and function assignment

Earlier, we saw that we can add a function dynamically to a Node. We can, of course, also do this via the Set method

acme$Set(cost = c(function(self) sum(sapply(self$children, 
                                            function(child) GetAttribute(child, "cost")))), 
         filterFun = isNotLeaf)
print(acme, "cost")

##                           levelName    cost
## 1  Acme Inc.                        4950000
## 2   ¦--Accounting                   1500000
## 3   ¦   ¦--New Software             1000000
## 4   ¦   °--New Accounting Standards  500000
## 5   ¦--Research                     2750000
## 6   ¦   ¦--New Product Line         2000000
## 7   ¦   °--New Labs                  750000
## 8   °--IT                            700000
## 9       ¦--Outsource                 400000
## 10      ¦--Go agile                  250000
## 11      °--Switch to R                50000

acme$IT$AddChild("Paperless", cost = 240000)
print(acme, "cost")

##                           levelName    cost
## 1  Acme Inc.                        5190000
## 2   ¦--Accounting                   1500000
## 3   ¦   ¦--New Software             1000000
## 4   ¦   °--New Accounting Standards  500000
## 5   ¦--Research                     2750000
## 6   ¦   ¦--New Product Line         2000000
## 7   ¦   °--New Labs                  750000
## 8   °--IT                            940000
## 9       ¦--Outsource                 400000
## 10      ¦--Go agile                  250000
## 11      ¦--Switch to R                50000
## 12      °--Paperless                 240000