DiagrammeR (version 0.9.0)

trav_both: Traverse from one or more selected nodes onto neighboring nodes

Description

From a graph object of class dgr_graph move from one or more nodes present in a selection to other nodes that are connected by edges, replacing the current nodes in the selection with those nodes traversed to. An optional filter by node attribute can limit the set of nodes traversed to.

Usage

trav_both(graph, conditions = NULL, copy_attrs_from = NULL, agg = "sum")

Arguments

graph

a graph object of class dgr_graph.

conditions

an option to use filtering conditions for the traversal.

copy_attrs_from

providing a node attribute name will copy those node attribute values to the traversed nodes. Any values extant on the nodes traversed to will be replaced.

agg

if a node attribute is provided to copy_attrs_from, then an aggregation function is required since there may be cases where multiple edge attribute values will be passed onto the traversed node(s). To pass only a single value, the following aggregation functions can be used: sum, min, max, mean, or median.

Value

a graph object of class dgr_graph.

Examples

Run this code
# NOT RUN {
# Set a seed
set.seed(23)

# Create a simple graph
graph <-
  create_graph() %>%
  add_n_nodes(
    2, type = "a",
    label = c("asd", "iekd")) %>%
  add_n_nodes(
    3, type = "b",
    label = c("idj", "edl", "ohd")) %>%
  add_edges_w_string(
    "1->2 1->3 2->4 2->5 3->5",
    rel = c(NA, "A", "B", "C", "D"))

# Create a data frame with node ID values
# representing the graph edges (with `from`
# and `to` columns), and, a set of numeric values
df_edges <-
  data.frame(
    from = c(1, 1, 2, 2, 3),
    to = c(2, 3, 4, 5, 5),
    values = round(rnorm(5, 5), 2))

# Create a data frame with node ID values
# representing the graph nodes (with the `id`
# columns), and, a set of numeric values
df_nodes <-
  data.frame(
    id = 1:5,
    values = round(rnorm(5, 7), 2))

# Join the data frame to the graph's internal
# edge data frame (edf)
graph <-
  graph %>%
  join_edge_attrs(df_edges) %>%
  join_node_attrs(df_nodes)

get_node_df(graph)
#>   id type label values
#> 1  1    a   asd   8.58
#> 2  2    a  iekd   7.22
#> 3  3    b   idj   5.95
#> 4  4    b   edl   6.71
#> 5  5    b   ohd   7.48

get_edge_df(graph)
#>   id from to  rel values
#> 1  1    1  2 <NA>   6.00
#> 2  2    1  3    A   6.11
#> 3  3    2  4    B   4.72
#> 4  4    2  5    C   6.02
#> 5  5    3  5    D   5.05

# Perform a simple traversal from node `3`
# to adjacent nodes with no conditions on
# the nodes traversed to
graph %>%
  select_nodes_by_id(3) %>%
  trav_both() %>%
  get_selection()
#> [1] 1 5

# Traverse from node `2` to any adjacent
# nodes, filtering to those nodes that have
# numeric values less than `8.0` for
# the `values` node attribute
graph %>%
  select_nodes_by_id(2) %>%
  trav_both(
    conditions = "values < 8.0") %>%
  get_selection()
#> [1] 4 5

# Traverse from node `5` to any adjacent
# nodes, filtering to those nodes that
# have a `type` attribute of `b`
graph %>%
  select_nodes_by_id(5) %>%
  trav_both(
    conditions = "type == 'b'") %>%
  get_selection()
#> [1] 3

# Traverse from node `2` to any adjacent
# nodes, and use multiple conditions for the
# traversal (using a vector in `conditions`
# creates a set of `AND` conditions)
graph %>%
  select_nodes_by_id(2) %>%
  trav_both(
    conditions = c(
      "type == 'a'",
      "values > 8.0")) %>%
  get_selection()
#> [1] 1

# Traverse from node `2` to any adjacent
# nodes, and use multiple conditions with
# a single-length vector (here, using a
# `|` to create a set of `OR` conditions)
graph %>%
  select_nodes_by_id(2) %>%
  trav_both(
    conditions = c(
      "type == 'a' | values > 8.0")) %>%
  get_selection()
#> [1] 1

# Traverse from node `2` to any adjacent
# nodes, and use a regular expression as
# a filtering condition
graph %>%
  select_nodes_by_id(2) %>%
  trav_both(
    conditions = "grepl('..d', label)") %>%
  get_selection()
#> [1] 1 5

# Create another simple graph to demonstrate
# copying of node attribute values to traversed
# nodes
graph <-
  create_graph() %>%
  add_path(5) %>%
  select_nodes_by_id(c(2, 4)) %>%
  set_node_attrs_ws("value", 5)

# Show the graph's internal node data frame
graph %>% get_node_df()
#>   id type label value
#> 1  1 <NA>     1    NA
#> 2  2 <NA>     2     5
#> 3  3 <NA>     3    NA
#> 4  4 <NA>     4     5
#> 5  5 <NA>     5    NA

# Show the graph's internal edge data frame
graph %>% get_edge_df()
#>   id from to  rel
#> 1  1    1  2 <NA>
#> 2  2    2  3 <NA>
#> 3  3    3  4 <NA>
#> 4  4    4  5 <NA>

# Perform a traversal from the inner nodes
# (`2` and `4`) to their adjacent nodes (`1`,
# `3`, and `5`) while also applying the node
# attribute `value` to target nodes; node `3`
# will obtain a `value` of 10 since a traversal
# to `3` will occur from `2` and `4` (and
# multiple values passed will be summed)
graph <-
  graph %>%
  trav_both(
    copy_attrs_from = "value",
    agg = "sum")

# Show the graph's internal node data frame
# after this change
graph %>% get_node_df()
#>   id type label value
#> 1  1 <NA>     1     5
#> 2  2 <NA>     2     5
#> 3  3 <NA>     3    10
#> 4  4 <NA>     4     5
#> 5  5 <NA>     5     5
# }

Run the code above in your browser using DataCamp Workspace