Unexpected behavior with a string stored in a variable in PowerShell

Question

I'm getting some odd behavior from Excel's Cells.Find() method:

Variable I'm searching on:

PS > $volumename
vol_01       

PS > $volumename.GetType()

IsPublic IsSerial Name                                     BaseType                                                                                                                                            
-------- -------- ----                                     --------                                                                                                                                            
True     True     String                                   System.Object

produces no results:

PS > $sheet.Cells.Find($volumename).Row

but if I manually copy and paste the value of that variable:

PS > $volumename = "vol_01"
PS > $volumename.GetType()

IsPublic IsSerial Name                                     BaseType                                                                                                                                            
-------- -------- ----                                     --------                                                                                                                                            
True     True     String                                   System.Object

Gets the value I am expecting:

PS > $sheet.Cells.Find($volumename).Row
198

They appear to be exactly the same type in every way to me. This doesn't happen for every case. Some volume names passthrough fine while others do not. I did scrub the volume name for this post as it has a customers naming convention. It is the same format as above and the same format as the volume names that work.

How did you manage opening a powershell terminal within an Excel sheet? — Dominique, Oct 04 '18 at 13:18
Use `[int[]] [char[]] 'hello' | % { '0x{0:x} {1}' -f $_, [char] $_ }` to inspect the string for hidden control characters. — mklement0, Oct 04 '18 at 13:25
@mklement0, Interesting. I am seeing some extraneous invisible characters at the end of the string with hex code 0xa0. How would I trim that off? string replace? — Nicholas Elliott, Oct 04 '18 at 13:31

mklement0 · Accepted Answer · 2022-12-07T00:01:41.573

The following snippet can be used to inspect a string for hidden control characters:

PS> & { [int[]] [char[]] $Args[0] | 
      % { '0x{0:x} [{1}]' -f $_, [char] $_ } } "vol_01`n"

0x76 [v]
0x6f [o]
0x6c [l]
0x5f [_]
0x30 [0]
0x31 [1]
0xa [
]

The first column is each character's Unicode code point ("ASCII code"), and the second column the character itself, enclosed in [...]

Note that I've added "`n" at the end of the string - a newline character (U+000A) - whose code point expressed as a hex. number is 0xa.

If, as in your case, the only unwanted part of the string is trailing whitespace, you can remove it as follows:

$volumename.TrimEnd() # trim trailing whitespace

In your case, the trailing whitespace is 0xa0, the NO-BREAK SPACE (U+00A0), which .TrimEnd() also removes, as Tom Blodget points out.

Simple function wrapper based on the above, for use with pipeline input:

filter debug-Chars { [int[]] [char[]] $_ | % { '0x{0:x} [{1}]' -f $_, [char] $_ } }

Sample use:

"vol_01`n" | debug-Chars

A more sophisticated function, named Debug-String, is available as an MIT-licensed Gist: It uses PowerShell's own escape-sequence notation to represent control characters in the input string, and also visualizes spaces as · by default; adding -UnicodeEscapes uses escape sequences for all non-ASCII-range characters.

Here's an example that first downloads and defines the function:

# Download and define the Debug-String function.
# NOTE: 
#  I can personally assure you that doing this is safe, but you
#  you should always check the source code first.
irm https://gist.github.com/mklement0/7f2f1e13ac9c2afaf0a0906d08b392d1/raw/Debug-String.ps1 | iex

"A b`nThis is a hyphen: '-', this is an en dash '–'" | Debug-String -SingleLine

Output:

The codepoint in question is U+00A0 NO-BREAK SPACE. .Trim() and TrimEnd() do remove it, too. Also, a PowerShell char is a UTF-16 code unit, not a Unicode codepoint (just like in the rest of .NET and many other environments since VB4 and Java). U+00A0 needs one UTF-16 code unit `[char]0x00a0` — Tom Blodget, Oct 05 '18 at 22:36
Thanks for the NO-BREAK SPACE correction, @TomBlodget - I've updated the answer accordingly. As for code point vs. code unit. True, PowerShell chars. are UTF-16 code units, but every code unit is also a Unicode code point. But not every code point by itself is a Unicode character (and thus not every .NET character is necessarily a full Unicode character). — mklement0, Oct 05 '18 at 22:48

Unexpected behavior with a string stored in a variable in PowerShell

1 Answers1

Linked