Delete all lines which contain a duplicate word - powershell

Question

I want to delete all the lines that contain one string and keep just the last line.
Eg:
a 1
a 2
a 3
b 1
b 2
I want to delete: a 1 a 2 b 1 and keep only the last lines: a 3 b 2.
I have tried something in powershell but witout success:
gc 1.txt | sort | get-unique

Answer 1

Assuming that you want to:
consider lines that share the same word at the start (a or b in your example) as group,
and return the last line from each such group,
use the Group-Object cmdlet:
Get-Content 1.txt | Group-Object { (-split $_)[0] } | ForEach-Object { $_.Group[-1] }
{ (-split $_)[0] } uses a property expression, via a script block ({ ... }, rather than a property name as the grouping criterion.
-split $_ splits each input line ($_) into an array of substrings by whitespace.
(...)[0] extracts the 1st token, i.e. the first whitespace-separated token on the line (a or b, in your sample data)
As for what you tried (showing your command with aliases expanded):
Get-Content 1.txt | Sort-Object | Get-Unique
Your Sort-Object and Get-Unique calls both operate on the full lines, which is not your intent: since all lines are unique when considered in full, they are all output.
Note that Sort-Object has a -Unique switch, so the following would come closer to what you want, but it wouldn't allow you to control which of the lines that share the same first word to return:
# !! INCORRECT, because you don't control which of the duplicates
# !! is returned, given that sorting is based on only the *first* word
# !! on each line.
PS> Get-Content 1.txt | Sort-Object { (-split $_)[0] } -Unique
a 1
b 1