Writing the same zip file from 0 bytes for multiple times

[sumtec]

When giving a large number of files from different directories as the paths argument, it seems to be writing the file over and over again in one execution.

For example, there are lots of directories under c:\logs, like:
C:\Logs\1
C:\Logs\1\This
C:\Logs\1\That
C:\Logs\2
C:\Logs\2\This
C:\Logs\2\That
C:\Logs\3
C:\Logs\3\This
C:\Logs\3\That
...

Each of these directories has many files. If there are 10k files and 1GB data, you will notice that the output zip file will show the size in such a manner:

...
103MB
15MB
75MB
121MB
0MB
30MB
147MB
26MB
...

It seems obvious to me that it's going to compress in such a way:

1. Compress some of the files in one directory and saved into a zip file.
2. Open the zip file and try to read it out, delete or truncate the old file, and repeat step 1.

It will take an unacceptable time to finish. Here is the script I used:
$files = ls "C:\Logs*" -Recurse | where { $_.Name -notlike '*.cab'} | select -ExpandProperty FullName

Compress-Archive $files -DestinationPath "C:\Compressed\Logs.zip" -CompressionLevel Optimal

[ayousuf23]

Can you provide more info to replicate this bug?