Adding vertex names when a named and unnamed graph are combined?

[schochastics]

See discussion in #1908 and #1932 (also related: #60)

What should be the strategy when a named and unnamed graph are combined by whatever means
(e.g. disjoint_union(), a graph product, "+",...)

I see two options:

1. throw an error (consistent with throwing an error when duplicated names are found)
2. create generic vertex names. An example implementation is at the bottom.

Currently, names are set to NA, which is in my opinion not desirable.

library(igraph)
library(purrr)

consolidate_names <- function(graphs, prefix = "V") {
  have_names <- map_lgl(graphs, function(g) "name" %in% vertex_attr_names(g))
  if (all(have_names)) {
    all_names <- list_c(map(graphs, function(g) V(g)$name))

    if (any(duplicated(all_names))) {
      cli::cli_abort("Duplicated vertex names found across graphs.")
    }
    return(graphs)
  }

  existing_names <- list_c(map(graphs, function(g) V(g)$name))

  # Create a counter for generating new names (check if named graph already has generic names)
  name_counter <- if (any(grepl(paste0(prefix, "[0-9]+"), existing_names))) {
    max(as.integer(gsub(paste0(prefix, "([0-9]+)"), "\\1", existing_names)), na.rm = TRUE) + 1
  } else {
    1
  }

  for (i in seq_along(graphs)) {
    if (!have_names[i]) {
      n <- vcount(graphs[[i]])
      num_id <- seq(name_counter, length.out = n)
      V(graphs[[i]])$name <- paste0(prefix, num_id)
      name_counter <- name_counter + n + 1
    }
  }

  return(graphs)
}

g1 <- g2 <- make_ring(4)
V(g1)$name <- c("A", "B", "C", "D")
graphs <- list(g1, g2)
consolidate_names(graphs)
#> [[1]]
#> IGRAPH 467efa4 UN-- 4 4 -- Ring graph
#> + attr: name (g/c), mutual (g/l), circular (g/l), name (v/c)
#> + edges from 467efa4 (vertex names):
#> [1] A--B B--C C--D A--D
#> 
#> [[2]]
#> IGRAPH 467efa4 UN-- 4 4 -- Ring graph
#> + attr: name (g/c), mutual (g/l), circular (g/l), name (v/c)
#> + edges from 467efa4 (vertex names):
#> [1] V1--V2 V2--V3 V3--V4 V1--V4

V(g1)$name <- c("V1", "V2", "V3", "V4")
graphs <- list(g1, g2)
consolidate_names(graphs)
#> [[1]]
#> IGRAPH 467efa4 UN-- 4 4 -- Ring graph
#> + attr: name (g/c), mutual (g/l), circular (g/l), name (v/c)
#> + edges from 467efa4 (vertex names):
#> [1] V1--V2 V2--V3 V3--V4 V1--V4
#> 
#> [[2]]
#> IGRAPH 467efa4 UN-- 4 4 -- Ring graph
#> + attr: name (g/c), mutual (g/l), circular (g/l), name (v/c)
#> + edges from 467efa4 (vertex names):
#> [1] V5--V6 V6--V7 V7--V8 V5--V8

^{Created on 2025-07-04 with reprex v2.1.1}

[szhorvat]

I believe in many scenarios indices are converted to names when necessary. So names become "1", "2", "3", ...

[schochastics]

Can you point me to one? In the combining graph scenario I could only find the NA case

[szhorvat]

I don't seem to be able to find this right now, but I have a distinct impression that I've seen vertex indices stringified at some point. It may not have been a direct conversion to a name attribute, just a context where strings-based names expected and the integer indices became strings.

I'll let you know if I find it.