In R - Substring based on repeating character - r

Question

I have two tables. In one table(IPTable), there is a column in one table containing IP addresses (which look like this: "10.100.20.13"). I am trying to match each of those to the data in a column in the other table (SubnetTable) which holds subnet addresses (which look like this: "10.100.20", essentially a shortened version of the IP address - everything before the 3rd period). Both variables appear to be chr vectors.
Essentially the raw IP data looks like this:
IPTable$IPAddress
10.100.20.13
10.100.20.256
10.100.200.23
101.10.13.43
101.100.200.1
and the raw Subnet data I am comparing it against looks like this:
SubnetTable$Subnet
Varies
10.100.20
Remote Subnet
10.100.200, 101.10.13
Unknown Subnet
Notes:
sometimes the subnet entries contain two subnets within a field separated by a comma
the IPAddress field doesn't have a consistent placement between the groups (e.g. - there could exist "10.110.20.13" as well as "101.10.20.13")
In a different scripting application I am able to simply compare these as strings in a foreach loop. In this logic, it loops through each entry in the Subnet data(SubnetTable), splits it against the comma (to account for the entries with multiple subnet addresses) and then checks to see if it finds a match in the IP Address field (e.g. - is "10.100.20" found anywhere in "10.100.20.13"). I use that field for a join/merge. In using R I understand that foreach looping isn't the most efficient way I should this and in the other application it takes a long time which is part of the reason I am moving to R.
I didn't see a method of doing the same thing against this type of data (I have done merges and joins but I don't see a way of doing that without getting to two variables alike enough to use to link the two tables).
In the past I have been able to use R methods like sqldf, charindex and leftstr to look for a particular character "." and pull everything before it but the difficulty here is that to do it that way, I need to look for the 3rd occurance of the period "." instead of the first. I didn't see a way of doing that but if there is a way, that might be best.
My next attempt was to use strsplit and sapply on the IP address with the idea of reassembling only the first three portions to create a subnet to match against (in a new column/variable). That looked like this:
IPClassC <- sapply(strsplit(Encrypt_Remaining5$IPAddress, "[.]"), `[`)
This gives a "Large List" which makes the data look like this:
chr [1:4] "10" "100" "20 "13"
But when attempting to put it back together I am also losing the period between the octets. Sample code:
paste(c(IPClassC[[1]][1:3]), sep ="[.]", collapse = "")
This produces something like this:
"1010020"
In the end I have two questions:
1) Is there a method for doing the easy comparison I did earlier (essentially doing a merge from the subnet variable of Table1 to "most" of the IP Address of Table2 basing it off of everything before the third period (".") without having to split it out and reassemble the IPAddress field?
2) If not, am I on the right track with trying to split and then reassemble? If so, what am I doing wrong with the reassembly or is there an easier/better way of doing this?
Thanks and let me know what else you need.

Answer 1

I think what you’re essentially asking is how to join these two tables, right? If this is the case, I would do it like this:
library(tidyr)
suppressPackageStartupMessages(library(dplyr))
IPTable <-
data.frame(
IPAddress =
c(
"10.100.20.13",
"10.100.20.256",
"10.100.200.23",
"101.10.13.43",
"101.100.200.1"
),
stringsAsFactors = FALSE
)
I am not sure, whether your SubnetTable really looks like this, i.e. mixing subnet addresses with other text? Anyway, this solution essentially ignores the other text.
SubnetTable <-
data.frame(
subnet_id = 1:5,
Subnet =
c(
"Varies",
"10.100.20",
"Remote Subnet",
"10.100.200, 101.10.13",
"Unknown Subnet"
),
stringsAsFactors = FALSE
)
First we separate multiple subnets into multiple rows. Note that this assumes that the SubnetTable$Subnet vector only contains a ", " to separate two subnets. I.e. there are no strings like this "Unknown, Subnet", or else these will be separated into two rows as well.
SubnetTable_tidy <- tidyr::separate_rows(SubnetTable, Subnet, sep = ", ")
SubnetTable_tidy
#> subnet_id Subnet
#> 1 1 Varies
#> 2 2 10.100.20
#> 3 3 Remote Subnet
#> 4 4 10.100.200
#> 5 4 101.10.13
#> 6 5 Unknown Subnet
Next we extract the Subnet by replacing/deleting a dot (\\.) followed by one to three numbers (\\d{1,3}) followed by the end of the string ($) from IPTable$IPAddress.
IPTable$Subnet <- gsub("\\.\\d{1,3}$", "", IPTable$IPAddress)
IPTable
#> IPAddress Subnet
#> 1 10.100.20.13 10.100.20
#> 2 10.100.20.256 10.100.20
#> 3 10.100.200.23 10.100.200
#> 4 101.10.13.43 101.10.13
#> 5 101.100.200.1 101.100.200
Now we can join both tables.
IPTable_subnet <-
dplyr::left_join(
x = IPTable,
y = SubnetTable_tidy,
by = "Subnet"
)
IPTable_subnet
#> IPAddress Subnet subnet_id
#> 1 10.100.20.13 10.100.20 2
#> 2 10.100.20.256 10.100.20 2
#> 3 10.100.200.23 10.100.200 4
#> 4 101.10.13.43 101.10.13 4
#> 5 101.100.200.1 101.100.200 NA

Answer 2

unlist(strsplit(SubnetTable$Subnet,split=",")) %in%
gsub("^(\\d{2,3}.\\d{2,3}.\\d{2,3}).*$","\\1",IPTable$IPAddress)
This will give you a vector of class logical that matches a TRUE/FALSE to each item in Subnet (giving multiple responses for items with commas in them). Alternatively, you can flip the two sides to get a list of logicals for each of the IPAddress, telling you if it exists in the subnet list.
Is this what you were looking for?
You can also achieve a similar result with charmatch:
sapply(strsplit(SubnetTable$Subnet, split=","), charmatch, IPTable$IPAddress)
This gives the following result with your sample data:
[[1]]
[1] NA
[[2]]
[1] 0
[[3]]
[1] NA
[[4]]
[1] 3 NA
[[5]]
[1] NA
Note that when there is a single match, you get the index for it, but where there are multiple matches, the value is 0.
Finally, the flip of this one will give you a list of the indexes in subnet where the IPaddresses match to:
sapply(gsub("^(\\d{2,3}.\\d{2,3}.\\d{2,3}).*$","\\1",IPTable$IPAddress), charmatch, SubnetTable$Subnet)
results in:
10.100.20 10.100.20 10.100.200 101.10.13 101.100.200
2 2 4 NA NA