powershell extract number with regex [duplicate] - regex

Question

This question already has answers here:
How do I return only the matching regular expression when I select-string(grep) in PowerShell?
(8 answers)
Closed 4 years ago.
I need to extract only the numbers from external file. I am using the following command:
(Get-Content -Path .\log.html) | Select-String -Pattern 'load is'
It then returns:
<tr><td>server-67 load is: 0</td></tr>
<tr><td>server-68 load is: 5875</td></tr>
<tr><td>server-69 load is: 6077</td></tr>
<tr><td>server-70 load is: 6072</td></tr>
<tr><td>server-71 load is: 5846</td></tr>
<tr><td>server-72 load is: 1900</td></tr>
<tr><td>server-73 load is: 1900</td></tr>
I only need to extract the number portion.
How can I do it?

Answer 1

What distinguishes this question from the near-duplicate at "How do I return only the matching regular expression when I select-string(grep) in PowerShell?" is the desire to extract substrings of interest via surrounding in-line context not to be included in the match:
PS> Select-String '(?<=load is: )\d+' .\log.html | ForEach-Object { $_.Matches[0].Value }
0
5875
6077
6072
5846
1900
1900
If you want to output actual numbers, simply place [int] (for instance) before $_.Matches[0].Value to cast (convert) the text results to an integer.
Select-String can accept file paths directly, so for a single file or a group of files matched by a wildcard expression you generally don't need to pipe from Get-Content.
(For processing entire directory subtrees, pipe from Get-ChildItem -File -Recurse).
Regex '(?<=load is: )\d+' uses a (positive) lookbehind assertion ((?<=...)) to match part of each line without including what was matched in the result; only the \d+ part - a nonempty run of digits - is captured.
Select-String outputs [Microsoft.PowerShell.Commands.MatchInfo] instances whose .Matches property contains the results of regex matching operation; its .Value property contains what the regex captured.
In the case at hand, the lookbehind solution is probably simplest, but an alternative solution is to use a capture group, which is ultimately more flexible:
# Same output as above.
Select-String 'load is: (\d+)' .\log.html | ForEach-Object {$_.Matches[0].Groups[1].Value}
What the capture group (the parenthesized subexpression, (...)) matched is available on the output objects' .Matches.Groups collection, whose element at index 0 contains the overall match, and element 1 containing the 1st capture groups, 2 the 2nd, and so on.

Answer 2

Here's one possibility:
(Get-Content -Path .\log.html) |
Where-Object {$_ -match '^<tr><td>server-(?<Server>\d{1,}) load is: (?<load>\d{1,})</td></tr>$'} |
ForEach-Object {
[PsCustomObject]#{"ServerNumber"=$matches.Server;"ServerLoad"=$matches.Load}
}
This will give you output like this:
ServerNumber ServerLoad
------------ ----------
67 0
68 5875
69 6077
70 6072
71 5846
72 1900
73 1900