combining data structures

Group

It's unlikely I'm trying this the best way, but I'm trying to create a 
data structure from elements like

nNode = 2
nn = vector("list",nNode)

nn[[1]] =  list(Node = "1", Connect.up = c(NULL), Connect.down = c(2,3))
nn[[2]] =  list(Node = "2", Connect.up = c(1), Connect.down = c(4,5))
....  #( and eventually many more nodes)

NodeList = as.data.frame(nn[[1]])
for(i in 2:nNode) {
   NodeList = rbind(NodeList,as.data.frame(nn[[i]]))
}

and is trying to create a data frame with many rows and three columns: 
Node, Connect.up,Connect.down
in which the Connect.up and Connect.down columns may be single numbers 
or vectors of numbers.  The above approach gives an error:

Error in data.frame(Node = "1", Connect.up = NULL, Connect.down = c(2,  :
   arguments imply differing number of rows: 1, 0, 2

My earlier try by brute force worked fine:

NodeList = as.data.frame(rbind(nn[[1]],nn[[2]]))

 > NodeList

   Node Connect.up Connect.down
1    1       NULL         2, 3
2    2          1         4, 5

and gives me what I want (many more rows eventually). But I want to do 
this generically from the problem context in a procedure so I won't know 
up front how many nodes I'll have.

Clearly I'm not understanding how referencing works for lists like I've 
created.  Can anyone shed light on this?

-- 
David K Stevens, P.E., Ph.D., Professor
Civil and Environmental Engineering
Utah Water Research Laboratory
8200 Old Main Hill
Logan, UT  84322-8200
435 797 3229 - voice
435 797 1363 - fax
david.stevens@




	[[alternative HTML version deleted]]

Thanks for the reply. Two things - I must have something missing because 
copying and pasting your example gave me an error

... your definitions of nn

  Output = do.call(as.data.frame(rbind),nn)

Error in as.data.frame.default(rbind) :
   cannot coerce class '"function"' into a data.frame

The second is the need, and why I'm not sure this is the best way. This 
is to customize the flow charting library code from the library 
'diagram' to use a node-and-link characterization of a network - think 
PERT charts or perhaps linking river segments in a water quality model. 
Each Node will be a, say, circle and has attributes of being connected 
up to one or more upstream nodes and down to one or more downstream 
nodes. So the Connect.up column is a vector up upstream connections, 
usually one but sometimes > one, and likewise Connect.down. There is an 
accompanying table of links with attributes of which nodes are at each 
end of a link and other metadata that describe the link (e.g. the length 
of time required to traverse the link, its name etc. That said, my 
thought was that the situation was too simple to fire up a full-blown 
object system beyond what R provides natively. I guess it's like making 
a data frame that has some 3-d elements.



On 2/3/2012 9:32 PM, Pete Brecknock wrote:

nn=list()

nn[[1]] =  list(Node = "1", Connect.up = c(NULL), Connect.down = c(2,3))
nn[[2]] =  list(Node = "2", Connect.up = c(1), Connect.down = c(4,5))
nn[[3]] =  list(Node = "3", Connect.up = c(NULL), Connect.down = c(2,3))
nn[[4]] =  list(Node = "4", Connect.up = c(1), Connect.down = c(4,5))

Output = do.call(as.data.frame(rbind),nn)

-- 
David K Stevens, P.E., Ph.D., Professor
Civil and Environmental Engineering
Utah Water Research Laboratory
8200 Old Main Hill
Logan, UT  84322-8200
435 797 3229 - voice
435 797 1363 - fax
david.stevens@




	[[alternative HTML version deleted]]