network (version 1.13.0)

attribute.methods: Attribute Interface Methods for the Network Class

Description

These methods get, set, list, and delete attributes at the network, edge, and vertex level.

Usage

delete.edge.attribute(x, attrname)
delete.network.attribute(x, attrname)
delete.vertex.attribute(x, attrname)
get.edge.attribute(el, attrname, unlist = TRUE, na.omit = FALSE, null.na = FALSE, deleted.edges.omit = FALSE)
get.edge.value(x, attrname, unlist = TRUE, na.omit = FALSE, null.na = FALSE, deleted.edges.omit = FALSE)
get.network.attribute(x, attrname, unlist = FALSE)
get.vertex.attribute(x, attrname, na.omit = FALSE, null.na = TRUE, unlist = TRUE)
network.vertex.names(x)
list.network.attributes(x)
list.edge.attributes(x)
list.vertex.attributes(x)
set.edge.attribute(x, attrname, value, e=seq_along(x$mel))
set.edge.value(x, attrname, value, e=seq_along(x$mel))
set.network.attribute(x, attrname, value)
set.vertex.attribute(x, attrname, value, v=seq_len(network.size(x)))
network.vertex.names(x) <- value

Arguments

el

a list of edges (possibly network$mel), or an object of class network from which the full list of edges will be extracted

x

an object of class network.

attrname

the name of the attribute to get or set.

unlist

logical; should retrieved attribute values be unlisted prior to being returned?

na.omit

logical; should retrieved attribute values corresponding to vertices/edges marked as 'missing' be removed?

deleted.edges.omit

logical: should the elements corresponding to deleted edges be removed?

null.na

logical; should NULL values (corresponding to vertices or edges with no values set for the attribute) be replaced with NAs in output?

value

values of the attribute to be set; these should be in vector or list form for the edge and vertex cases, or matrix form for set.edge.value.

e

IDs for the edges whose attributes are to be altered.

v

IDs for the vertices whose attributes are to be altered.

Value

For the list.attributes methods, a vector containing attribute names. For the get.attribute methods, a list containing the values of the attribute in question (or simply the value itself, for get.network.attribute). For the set.attribute and delete.attribute methods, a pointer to the updated network object.

Details

The list.attributes functions return the names of all edge, network, or vertex attributes (respectively) in the network. All attributes need not be defined for all elements; the union of all extant attributes for the respective element type is returned.

The get.attribute functions look for an edge, network, or vertex attribute (respectively) with the name attrname, returning its values. Note that, to retrieve an edge attribute from all edges within a network x, x$mel should be used as the first argument to get.edge.attribute; get.edge.value is a convenience function which does this automatically. As of v1.7.2, if a network object is passed to get.edge.attribute it will automatically call get.edge.value instead of returning NULL. When the parameters na.omit, or deleted.edges.omit are used, the position index of the attribute values returned will not correspond to the vertex/edge id. To preserved backward compatibility, if the edge attribute attrname does not exist for any edge, get.edge.attribute will still return NULL even if null.na=TRUE

network.vertex.names is a convenience function to extract the "vertex.names" attribute from all vertices.

The set.attribute functions allow one to set the values of edge, network, or vertex attributes. set.edge.value is a convenience function which allows edge attributes to be given in adjacency matrix form, and the assignment form of network.vertex.names is likewise a convenient front-end to set.vertex.attribute for vertex names. The delete.attribute functions, by contrast, remove the named attribute from the network, from all edges, or from all vertices (as appropriate). If attrname is a vector of attribute names, each will be removed in turn. These functions modify their arguments in place, although a pointer to the modified object is also (invisibly) returned.

Additional practical example of how to load and attach attributes are on the loading.attributes page.

Some attribute assignment/extraction can be performed conveniently through the various extraction/replacement operators, although they may be less efficient. See the associated man page for details.

References

Butts, C. T. (2008). “network: a Package for Managing Relational Data in R.” Journal of Statistical Software, 24(2). http://www.jstatsoft.org/v24/i02/

See Also

loading.attributes,network, as.network.matrix, as.sociomatrix, as.matrix.network, network.extraction

Examples

#Create a network with three edges
m<-matrix(0,3,3)
m[1,2]<-1; m[2,3]<-1; m[3,1]<-1
g<-network(m)

#Create a matrix of values corresponding to edges
mm<-m
mm[1,2]<-7; mm[2,3]<-4; mm[3,1]<-2

#Assign some attributes
set.edge.attribute(g,"myeval",3:5)
set.edge.value(g,"myeval2",mm)
set.network.attribute(g,"mygval","boo")
set.vertex.attribute(g,"myvval",letters[1:3])
network.vertex.names(g) <- LETTERS[1:10]

#List the attributes
list.edge.attributes(g)
list.network.attributes(g)
list.vertex.attributes(g)

#Retrieve the attributes
get.edge.attribute(g$mel,"myeval")  #Note the first argument!
get.edge.value(g,"myeval")          #Another way to do this
get.edge.attribute(g$mel,"myeval2") 
get.network.attribute(g,"mygval")
get.vertex.attribute(g,"myvval")
network.vertex.names(g)

#Purge the attributes
delete.edge.attribute(g,"myeval")
delete.edge.attribute(g,"myeval2")
delete.network.attribute(g,"mygval")
delete.vertex.attribute(g,"myvval")

#Verify that the attributes are gone
list.edge.attributes(g)
list.network.attributes(g)
list.vertex.attributes(g)

#Note that we can do similar things using operators
g %n% "mygval" <- "boo"               #Set attributes, as above
g %v% "myvval" <- letters[1:3]
g %e% "myeval" <- mm
g[,,names.eval="myeval"] <- mm          #Another way to do this
g %n% "mygval"                        #Retrieve the attributes
g %v% "myvval"
g %e% "mevval"
as.sociomatrix(g,"myeval")              # Or like this