You can utilize the patternMatch function to split the device name string into different columns as follows:

parts(deviceName) =
  patternMatch(replace(deviceName, "_", " "), `{part1:string} {part2:string} {part3:string}`);

foreach device in network.devices

select {
    device: device.name,
    device1: parts(device.name)?.part1,
    device2: parts(device.name)?.part2,
    device3: parts(device.name)?.part3,
}

Underscores in the string are replaced by spaces, since spaces are used by the patternMatch function to parse the string.

Note the (?) after the parts(device.name) function is called. This allows for the condition where the output of the function is null.

Hi @pjptsujit ,

Starting in 24.10, NQE has regex matching functions. This gives you another way to solve the above problem, like this:

testDeviceNames = ="aaa_bbb_ccc", "1_22222_33"];

foreach deviceName in testDeviceNames
let match = match(deviceName, re`(?<part1>\w+)_(?<part2>\w+)_(?<part3>\w+)`)
select {
  device: deviceName,
  part1: match?.part1,
  part2: match?.part2,
  part3: match?.part3
}

The nice thing here is that you do not need to first replace underscores with space characters. A minor improvement here, but may be helpful on more complex examples as well.