On Mon, Jun 15, 2015 at 10:47 AM, Elena Tea Russo <[email protected]> wrote: > > Thank you too, Gabor. > > I can just guess the problems that a "fixed" indexing in a graph could lead > to. > I suppose neither edges and vertex selectors are then reliable after I modify > my graph?
Exactly. These are tied to a given graph, and the "modified" graph is a different graph. In fact the next version of igraph will give an error if you want to use a vertex/edge sequence with another graph. Gabor > I'll have to rethink a bit my code, but I luckily can manage my problem > throughout an adjacency matrix. > > Thanks again! > > Et Russo > > Il giorno 15/giu/2015, alle ore 15:20, Gábor Csárdi <[email protected]> > ha scritto: > >> Hi, >> >> I understand that reassigning ids is very difficult to work with, it >> was a decision long time ago, and unfortunately we have to live with >> it now, and for the foreseeable future. >> >> A useful way of thinking about this, especially in R, is that igraph >> _never_ modifies a graph. It always creates new graphs. So when you >> say delete.edges(), the old graph is not modified, and a new graph is >> created based on it. >> >> In general, you cannot assume anything about the relationship between >> the internal ids of the old graph and the new graph. >> >> I hope this helps, >> Gabor > > > _______________________________________________ > igraph-help mailing list > [email protected] > https://lists.nongnu.org/mailman/listinfo/igraph-help _______________________________________________ igraph-help mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/igraph-help
