sna (version 2.8)

write.nos: Write Output Graphs in (N)eo-(O)rg(S)tat Format

Description

Writes a graph stack to an output file in NOS format.

Usage

write.nos(x, file, row.col = NULL, col.col = NULL)

Value

None.

Arguments

x

a graph or graph stack (all graphs must be of common order).

file

string containing the output file name.

row.col

vector of row labels (or "row colors").

col.col

vector of column labels ("column colors").

Author

Carter T. Butts

Details

NOS format consists of three header lines, followed by a whitespace delimited stack of raw adjacency matrices; the format is not particularly elegant, but turns up in certain legacy applications (mostly at CMU). write.nos provides a quick and dirty way of writing files NOS, which can later be retrieved using read.nos.

The content of the NOS format is as follows:

<m>

<n> <o>

<kr1> <kr2> ... <krn> <kc1> <kc2> ... <kcn>

<a111> <a112> ... <a11o>

<a121> <a122> ... <a12o>

...

<a1n1> <a1n2> ... <a1no>

<a211> <a212> ... <a21o>

...

<a2n1> <a2n2> ... <a2no>

...

<amn1> <amn2> ... <amno>

where <abcd> is understood to be the value of the c->d edge in the bth graph of the file. (As one might expect, m, n, and o are the numbers of graphs (matrices), rows, and columns for the data, respectively.) The "k" line contains a list of row and column "colors", categorical variables associated with each row and column, respectively. Although originally intended to communicate exchangability information, these can be used for other purposes (though there are easier ways to deal with attribute data these days).

Note that NOS format only supports graph stacks of common order; graphs of different sizes cannot be stored within the same file.

See Also

read.nos, write.dl, write.table

Examples

if (FALSE) {
#Generate a random graph stack
g<-rgraph(5,10)

#This would save the graphs in NOS format
write.nos(g,file="testfile.nos")

#We can also read them back, like so:
g2<-read.nos("testfile.nos")
}