[Nauty] How to produce a histogram of graphs?

[Joost Winne]

> I am running some computer simulations in which I input Nauty a large list of graphs (up to 65,000 or so, and up to about 30 vertices per graph).  I then want to produce a histogram of the different isomorphims classes in that list.  Right now I am just having Nauty output the canonical form of all 65,000 graphs, and then I go through and sort them.  This is rather slow, since the canonical form is represented as a string.  I was wondering if it were possible to have Nauty either:
>
> 1) Output a numerical representation of each graph
> (i.e. the unique binary number produced by looking at the adjacency
> matrix of each graph), since I could then produce a binary search tree
> based on this number -- which is much faster than just comparing
> strings.

Is using the following also slow: (sorting a file uniquely)

sort -u graph6_File_With_One_Graph_On_Each_Line.g6 > unique_Graphs.g6

To count the number of graphs (lines in the file) just use:

wc -l file

Joost