String. Possible values are "edgelist"
and "gml"
.
If format
is "edgelist"
, we assume the following edgelist matrix format. Each row is assumed to be of the form (from_node_id
to_node_id
time_stamp
).
from_node_id
is the id of the source node. to_node_id
is the id of the destination node. time_stamp
is the arrival time of the edge. from_node_id
and to_node_id
are assumed to be integers that are at least \(0\). They need not to be contiguous.
To register a new node \(i\) at time \(t\) without any edge, add a row with format (i -1 t
). This works for both undirected and directed networks.
time_stamp
can be either numeric or string. The value of a time-stamp can be arbitrary, but we assume that a smaller time_stamp (regarded so by the sort
function in R
) represents an earlier arrival time. Examples of time-stamps that satisfy this assumption are the integer 0:T
, the string format `yyyy-mm-dd', and the POSIX time.
If format
is "gml"
, there must be a binary field directed
indicating the type of the network (0
: undirected, 1
: directed). The required fields for an edge are: source
, target
, and time
. source
and target
are the ID of the source node and the target node, respectively. time
is the time-stamp of the edge. The required fields for a node are: id
, isolated
(binary) and time
. The binary field isolated
indicates whether this node is an isolated node when it enters the system or not. If isolated
is 1
, then time
must contain the node's appearance time. If isolated
is 0
, then we can automatically infer the node's appearance time from its edges, so the field time
in this case can be NULL
. The assumptions on node IDs and the format of time-stamps are the same as in the case when format = "edgelist"
. See graph_to_file
to see detail on the format of the gml
file this package outputs.