I recently got a fun email from a client. Their tenant had run over its storage allocation. After some quick investigation they realized that they had a few dozen files with a lot of versions, the average number of versions was over 1200 per file. Some files had over 3000 versions. The kicker is that these files were big, dozens or hundreds of MBs each. Thousands of files, 100s of MBs a piece? Pretty soon you’re talking about some real storage space! His question to me was how he could easily delete these unneeded versions quickly and easily? PowerShell, of course!
Before we get to the golden PowerShell nuggets at the end, how did they end up here in the first place? Versions have been around forever in SharePoint but in the last couple of years there was a change. As part of their strategy to protect us from ransomware, Microsoft turned on versioning for all Document Libraries in SPO, and set the maximum number of versions to 500. That way if malware encrypted your site you could just roll your documents back to an unencrypted version. Pretty clever approach. The bad news is that if you don’t know this is in place, like my customer didn’t, you can chew up a lot of space by frequently uploading large files. Back to our story…
My mancrush on the PnP PowerShell is well documented, so of course that’s the first place I looked for a solution. Sure enough, there it was, shining like a beacon of hope on a foggy night, Remove-PnPFileVersion. Do you hear angels singing? I know I do. I sent that off to my customer, dusted off my hands, and leaned back, put my hands behind my head and basked in self satisfaction. It was glorious. Until the customer replied…
He didn’t want to delete all of the versions of a file. He wanted to something like delete all except the last 5 versions. Remove-PnPFileVersion has 2 parameter sets. One that deletes a single version (by its ID number) and one that deletes all versions. No middle ground. While my previous victory was short lived, I knew PowerShell would come through for me here too.
Long story short, I scribbled down a quick PowerShell script that will delete the file versions beyond the number you wish to keep. This customer had a CSV file of the files they wanted to prune, so I added support for that. I have posted the files on GitHub (DeleteOldVersions.ps1 and VersionDelete.csv ) and I’ll go over the mechanics here.
Here is what the CSV file looks like, if you want to use that:
The Remove-PnPFileVersion cmdlet natively handles URLs in the forms of lines 2-5. I added support for the URL having the tenant name as well because a report the customer had included that and I wanted to make it easy for them to read. I’m good like that.
Here’s the code:
First I connect to the site:
You’ll need to adjust those to your own situation. Then I load up the CSV file. If you don’t want to do this with a CSV file you don’t have to. You can manually put the file name in.
$FileList = Import-Csv .\VersionDelete.csv
Then I pick how many file versions I want to keep:
$VersionsToKeep = 5
Next I walk through the $FileList.
foreach($File in $FileList) {
If the FileName property has the $SiteUrl in it I take it out.
$Filename = $File.FileName.Replace($SiteUrl,"")
Then I grab all the FileVersions of the file:
$FileVersions = Get-PnPFileVersion -Url $Filename
Get-PnPFileVersion does not show the Current Version, so it will always show one fewer version than what you see in the UI. If the number of versions is greater than the version we said we wanted to keep in $VersionsToKeep then I create a list of versions in $DeleteVersionList to delete:
if ($FileVersions.Count -gt $VersionsToKeep) {
$DeleteVersionList = ($FileVersions[0..$($FileVersions.Count - $VersionsToKeep)])
With that list in hand I walk through it and run Remove-PnPFileVersion against it. In the code in GitHub I have commented out the line (Line 33 as of 9/14/21) that actually deletes the version. You’ll have to uncomment that to do anything.
foreach($VersionToDelete in $DeleteVersionList) {
Remove-PnPFileVersion -Url $Filename -Identity $VersionToDelete.Id –Force
To make piping easy I output the versions deleted as a CustomObject:
$Output = [PSCustomObject]@{
PSTypeName = 'TKDeletedFileVersion'
Filename = $Filename
DeletedVersion = $($VersionToDelete.VersionLabel)
}
$Output
When you look at the code you’ll notice a line at the top,
$VerbosePreference = "Continue"
If you uncomment that line it will light up the Write-Verbose statements in the code. Set it back to SilentyContinue to make them go away.
This is what it looks like with Verbosity off:
If you need to audit which versions are being deleted you can pipe that to Export-CSV and save it to a file.
If you want to create some versions to test this with you can use the following PowerShell:
$DocLib = “Shared Documents”
Get-PnPListItem -List $Doclib | select Id,@{l="FileLeafRef";e={$_.FieldValues.FileLeafRef}}
That’ll get you the files in the Document Library. You can use this command to touch them and create 6 new versions:
0..5 | foreach {Set-PnPListItem -List $Doclib -Identity 4 -Values @{"FileLeafRef"="Building materials licences to budget for Storytelling.docx"}}
It is also possible to filter versions by date. That way instead of deleting all but the last 5 versions, you would be able to delete all of the versions older than 30 days. That’s a blog post for a different day though.
tk
ShortURL: https://www.toddklindt.com/PoshDeleteSPOFileVersions