This vignette demonstrates some typical use cases for the ipaddress package.
Although IPv4 addresses are usually displayed in a human-readable format (e.g.
192.168.0.1), they are often saved to disk as integers to reduce storage requirements. R is unable to store the entire range of IPv4 addresses in its integer data type, but they can instead be stored in its numeric data type. More details are found under
Given this, it’s quite possible that you’ll receive IPv4 addresses represented as integers, and we’ll want to convert them to the
ip_address() vector type. Here’s an example of how to do that:
There are multiple equivalent ways to represent an IP network:
ip_network() function can handle the first 3 options, we use the
common_network() function for the final option.
tibble( start = ip_address(c("192.168.0.0", "2001:db8::")), end = ip_address(c("192.168.0.15", "2001:db8::ffff:ffff:ffff")) ) %>% mutate(network = common_network(start, end)) #> # A tibble: 2 x 3 #> start end network #> <ip_addr> <ip_addr> <ip_netwk> #> 1 192.168.0.0 192.168.0.15 192.168.0.0/28 #> 2 2001:db8:: 2001:db8::ffff:ffff:ffff 2001:db8::/80
Note that this approach assumes the two addresses do actually correspond to the first and last addresses of the network, otherwise an expanded network will be returned (see
help("common_network") for details).
A very common task is to check if an address is within a network, so the ipaddress package provides a couple of different functions to help with this workflow:
is_within_any(). We also provide
is_supernet() to test if a network is within another network.
To see how these functions can be used in practice, let’s consider a couple of IP networks:
and a handful of addresses:
First, we’ll check if each address is in any of our networks.
But what if we need to know which of our networks the address was found in? We can do that using the excellent fuzzyjoin package together with the
my_addresses %>% fuzzyjoin::fuzzy_left_join(my_networks, c("address" = "network"), is_within) #> # A tibble: 4 x 3 #> address network label #> <ip_addr> <ip_netwk> <chr> #> 1 192.168.100.1 192.168.0.0/16 Private #> 2 184.108.40.206 NA <NA> #> 3 2001:db8::8a2e:370:7334 2001:db8::/32 Documentation #> 4 ::1 NA <NA>
ipaddress provides functions to sample from a specific network (
sample_network()) or the entire address space (
sample_ipv6()). However, it can be more difficult to sample from the majority of address space, while excluding certain networks.
A good example is sampling public IPv4 addresses. The simplest solution is to use an accept-reject algorithm – sampling the entire IPv4 address space and rejecting addresses that are reserved.
We now sample 10 addresses and make sure they are all public addresses.
tibble(address = sample_public(10)) %>% mutate(is_public = is_global(address)) #> # A tibble: 10 x 2 #> address is_public #> <ip_addr> <lgl> #> 1 220.127.116.11 TRUE #> 2 18.104.22.168 TRUE #> 3 22.214.171.124 TRUE #> 4 126.96.36.199 TRUE #> 5 188.8.131.52 TRUE #> 6 184.108.40.206 TRUE #> 7 220.127.116.11 TRUE #> 8 18.104.22.168 TRUE #> 9 22.214.171.124 TRUE #> 10 126.96.36.199 TRUE