Split string in powershell with space-CodePudding

I have a string as "Device: ILO 5 Firmware Version : 2:33 Firmware date : 10-10-2020" I want to split this string as below in an array

ILO 5
2:33
10-10-2020

How can it be done in powershell?

CodePudding user response：

You could use a regular expression to match and extract the values:

$string = "Device: ILO 5 Firmware Version : 2:33 Firmware date : 10-10-2020"

if($string -match 'Device:\s*(.*?)\s*Firmware Version :\s*(.*?)\s*Firmware date :\s*([\d\-] )'){
  [pscustomobject]@{
    Device = $Matches[1]
    FWVersion = $Matches[2]
    FWDate = $Matches[3]
  }
}

\s* matches 0 or more spaces, .*? matches 0 or more of any characters, non-greedily, and [\d\-] will find any sequence of one or more characters that are either digits or -. The () around .*? and [\d\-] instructs the regex engine to capture the matching values, which is how we can then extract those values from the automatic $Matches variable afterwards.

This should produce an object with those three properties:

Device FWVersion FWDate
------ --------- ------
ILO 5  2:33      10-10-2020

CodePudding user response：

This might be ugly and there might be other ways but this is one way to do it:

First create an empty array:

$Array = @()

Get all the elements in a temp array:

$TempArray = "Device: ILO 5 Firmware Version : 2:33 Firmware date : 10-10-2020" -split(' ')

Next add the first values in the temparray, index 1 and 2, as a string:

$array  = $temparray[1..2] -as [string]

Then add the others:

$Array  = $temparray[6]
$Array  = $temparray[-1]

This adds the sixth and the last element (-1). The result is:

ILO 5
2:33
10-10-2020

CodePudding user response：

You can split the string on whitespace, but ILO and 5 will be two different array elements. The comma operator has a higher precedence than the join operator, hence the parentheses.

$string = 'Device: ILO 5 Firmware Version : 2:33 Firmware date : 10-10-2020'
$split = -split $string
($split[1,2] -join ' '),$split[6,10]

ILO 5
2:33
10-10-2020

CodePudding user response：

A generalized solution, which:

doesn't require knowing the field names in advance.
returns the field values only (as stipulated in your question - see bottom section if want to capture name-value pairs):

# Sample input string.
$str = 'Device: ILO 5 Firmware Version : 2:33 Firmware date : 10-10-2020'

# Split so as to only return field *values*.
$str -split '(?:^|  )\w  ?\w* *:  ' -ne ''

Assumptions:

No value contains substring ': '
Names contain only letters and _ and at most one space (i.e. are composed of one or two words).

Output:

ILO 5
2:33
10-10-2020

For an explanation of the above regex used with the -split operator, including the ability to experiment with it, see this regex101.com page.
-ne '' filters out the unwanted empty token that results from the regex also matching at the start of the input string.

To also capture the field names, more work is needed:

# Sample input string.
$str = 'Device: ILO 5 Firmware Version : 2:33 Firmware date : 10-10-2020'

# Split into field names and values.
# Note the capture group - (...) - around the sub-expresssion
# that matches the field name. 
# -split includes capture-group matches in the result.
$namesAndValues = $str -split '(?:^|  )(\w  ?\w*) *:  ' -ne ''

# Construct an ordered hashtable from the name-value pairs
# You can easily convert it to a [pscustomobject] with: [pscustomobject] $result
$result = [ordered] @{}
for ($i = 0; $i -lt $namesAndValues.Count; $i  = 2) {
  $result[$namesAndValues[$i]] = $namesAndValues[$i 1]
}

# Output the result:
$result

Output:

Name                           Value
----                           -----
Device                         ILO 5
Firmware Version               2:33
Firmware date                  10-10-2020