I am trying to Declare $GroupName array and add a couple statements for the array. I have an additional code from a different script and am trying to create an array for that script. I am trying to test the Row.Category = group and the Row.excluded = false. Then it will check the $groupname Array for duplicates. If no duplicates are found, populate $groupname.
I am need help just creating the array since I am a powershell novice.
$GroupName = #(column1 = "Domain")(column2 = "Account")
If ($_.category0 -eq "Group") -and ($_.Excluded -match "False")
"Account0", "Domain0" | sort Account0 -Unique
How do I improve the following code above to display the 2 columns in the new CSV.
This is one method to achieve what you're after, creating a new CSV file:
$Object = New-Object -TypeName PSObject -Property #{ Domain=''; Account='' }
$Object | Export-CSV $env:temp\MyCSV.csv -NoTypeInformation
We create a new PSObject called $Object and provide a hashtable object with the structure you require to the -Property parameter.
We then pass the PSObject down the pipeline to the Export-CSV cmdlet. Adding -NoTypeInformation on the end allows us to create a clean CSV file without additional information from PowerShell that other systems would not recognise.
However, I should note there are other ways to create/generate CSVs, especially if you're wanting to populate them with data.
Related
I have 2 PScustomObjects that i would like to combine into a single csv as follows:
$Output1 = [PSCustomObject]#{
Timestamp1 = (Get-Date)
InstanceName1 = $ProcessName1
Count1 = $Processes1.Count
Memory1 = $MEMLoad1
CPU1 = $CPULoad1
}
$Output2 = [PSCustomObject]#{
Timestamp2 = (Get-Date)
InstanceName2 = $ProcessName2
Count2 = $Processes2.Count
Memory2 = $MEMLoad2
CPU2 = $CPULoad2
}
The CSV should have the titles TimeStamp1, InstanceName1......TimeStamp2, InstanceName2....
This code runs in a loop and exports data each pass.
Is there a way to do this? Also is there a way to do this dynamically to avoid replicating large amounts of code if i am to export data on say 100 PsCustomObjects, maybe lookping through the input data for the object and the putting in in one object and passing that to the csv while dynamically changing titles?
I use the following line to export. I've tried -InputObject $Output1, $Output2 but that gives gibberish.
Export-Csv -InputObject $Output1 -path $Path1 -NoTypeInformation -Append -Force
The only solution i have so far is to export multiple CSV's but that gets bulky fast.
Any help is greatly appreciated!
The best solution is to ensure all the objects your script/function outputs use the same schema (that is, they should all share the same common property names).
If that's not possible, you can use Select-Object to ensure all possible properties are selected for all objects:
$Output1,$Output2 |Select-Object Count1,Count2,CPU1,CPU2,InstanceName1,InstanceName2,Memory1,Memory2,Timestamp1,Timestamp2 |Export-Csv ...
This is obviously not very practical, so if you have many different object schemas, you'll want to automate the discovery of all possible property names.
To do that, we can inspect the psobject hidden memberset:
$allTheObjectsToExport = $Output1,$Output2
$propertyNames = $allTheObjectsToExport |ForEach-Object {
# Discover all the object's properties
$_.psobject.Properties |ForEach-Object {
# Get each property name
$_.Name
}
} |Sort-Object -Unique
# Now the CSV will have a column for every single unique property name
$allTheObjectsToExport |Select-Object $propertyNames |Export-Csv -Path $Path1 -NoTypeInformation
I have created a PsCustomObject, when the variable is called is ISE, it reads a table of the relevant data. However, if I try to compare the PsCustomObject with another object, the PsCustomObject doesn't read correctly. I'd like to tell the script if any of the lines in the existing CSV match the PSCustomObject do not export the data to the CSV, in other words skip duplicate rows in the CSV file. The CSV may or may not have multiple rows.
$fileInfo = #(
[pscustomobject]#{
user_id = $user
studio = $studio
function = $Task
end_time_local = $creationTime
asin = $ASIN
variant = $variant
process_class_id = $processClass
}
)
$currentData = Import-Csv "$scansFolder\$fileName.csv"
if($fileInfo -ne $currentData){
$fileInfo | Export-Csv "$scansFolder\$fileName.csv" -Append -NoTypeInformation -Force
}
[pscustomobject] is a .NET reference type, so comparing two instances[1] with -eq will test for reference equality (identity), i.e. if the two instances are one and the same object[2] - which is obviously not the case in your scenario.
Assuming that the properties of your custom objects are instances of value types or strings (which appears to be the case), you can use Compare-Object to compare objects by their property values, with the ability to compare two collections:
$fileInfo = #(
[pscustomobject]#{
user_id = $user
studio = $studio
function = $Task
end_time_local = $creationTime
asin = $ASIN
variant = $variant
process_class_id = $processClass
}
)
# Get the property names.
# This assumes that the CSV data has (at least) the same
# set of properties (columns).
$propNames = $fileInfo[0].psobject.properties.Name
$currentData = Import-Csv "$scansFolder\$fileName.csv"
# Compare the $fileInfo custom object(s) to the custom objects read
# from the CSV file and only export those that are unique to the RHS ('=>')
# back to the file, i.e., those that don't match $fileInfo.
Compare-Object -Property $propNames $fileInfo $currentData |
Where-Object SideIndicator -eq '=>' | Select-Object InputObject |
Export-Csv "$scansFolder\$fileName.csv" -Append -NoTypeInformation -Force
[1] Import-Csv outputs [pscustomobject] instances too.
[2] See the Equality Comparison help topic (written for C#, but applies analogously to PowerShell's -eq operator).
I'm reading in a CSV file which contains 25,000 records, and am reading each column into a psobject. Here is what I have so far:
$file = Import-CSV .\server.csv
$tempobj = New-Object psobject -Property #{
'Name' = $file.Name
'Group' = $file.Group
}
When this is ran, I get the correct results I want, being that $file.Name contains all the server names, and $file.Group contains the groups for servers. However, my issue is that I need to edit the names of each server without interfering with the .Group. Here is an example of what a server name look like as is.
WindowsAuthServer #{wdk9870WIN}
I need to remove WindowsAuthServer #{ and WIN} from each server name, leaving only the server name left, or for this example, wdk9870.
I tried using the -replace function ($tempobj.Name -replace "WindowsAuthServer #{",""), but it requires that I save the results to a new array, which then messes up or removes .Group entirely
Is there a different way to go about doing this? I'm lost.
Suppose your server.csv looks like this:
"Name","Group"
"WindowsAuthServer #{wdk9870WIN}","Group1"
"WindowsAuthServer #{wdk9880WIN}","Group2"
"WindowsAuthServer #{wdk9890WIN}","Group1"
"WindowsAuthServer #{wdk9900WIN}","Group1"
And you want to change the values in the Name column only, then something like this would probably do it:
Import-Csv .\server.csv | ForEach-Object {
New-Object psobject -Property #{
'Name' = ([regex]'#\{(\w+)WIN\}').Match($_.Name).Groups[1].Value
'Group' = $_.Group
}
}
This will output:
Name Group
---- -----
wdk9870 Group1
wdk9880 Group2
wdk9890 Group1
wdk9900 Group1
If you want, you can simply pipe this info to the Export-Csv cmdlet to save as a new CSV file. For that, just append | Export-Csv -Path .\server_updated.csv -NoTypeInformation to the code.
Hope that helps
I find myself continually faced with the need to store mixed-type data in some kind of a structure for later lookup.
For a recent example, I am performing data migration and I will store the old UUID, new UUID, source environment, target environment, and schema for an unknown number of entries.
I have been meeting this need by creating an array and inserting System.Objects with NoteProperty members for each of the columns of data.
This strikes me as a very clumsy approach but I feel like I may be limited by Powershell's functionality. If I need to, for example, locate all entries that used a particular schema, I write a foreach loop that sticks each entry with a matching schema name in a whole new array that I can return. I would really like the ability to more easily search for all objects that contain a member matching a particular value, modify existing members, etc.
Is there a better built-in data structure that will suit my needs, or is creating a custom object the right thing to do?
For reference, I'm doing something like this to create my structure:
$objectArray= #();
foreach(thing to process){
$tempObj = New-Object System.Object;
$tempObj | Add-Member -MemberType NoteProperty -Name "membername" -Value xxxxx
....repeat for each member...
$objectArray += $tempObj
}
If I need to find something in it, I then have to:
$matchingObjs = #()
foreach ($obj in $objectArray){
if($obj.thing -eq value){$matchingObjs += $obj}
}
This really sucks and I know there has to be a more elegant way. I'm still fairly new to powershell so I don't know what utilities it has to help me. I'm using v5.
With PowerShell 3.0 you could use a [PSCustomObject], here's an article on the different object creation methods.
Also setting the array equal to the output of the foreach loop will be more efficient than repeatedly recreating an array with +=.
$objectArray = foreach ($item in $collection) {
[pscustomobject]#{
"membername" = "xxxxx"
}
}
The Where-Object cmdlet or the .where() method looks like what you need in your second loop.
$matchingObjs = $objectArray | Where-Object {$_.thing -eq "value"}
It also sounds like you could use Where-Object/.where() to filter the initial data and just create an object which matches what you are looking for. For example:
$matchingObjs = $InputData |
Where-Object {$_.thing -eq "value"} |
ForEach-Object {
[pscustomobject]#{
"membername" = xxxxx
}
}
If your data can be expressed as key value pairs, then a hashtable will be the most efficient, see about_Hash_Tables for more info.
There is no built-in way to do what you are asking. One way is to segment your data into separate hashtables so you can do easy lookups by a common key, say the ID.
# Create a hastable for the IDs
$ids = #{};
foreach(thing to process){
$ids.Add($uid, 'Value')
}
# Find the $uid exists
$keyExists = $ids.Keys -Contains $uid
# Find value of stored for $uid
$keyValue = $ids[$uid]
As a side note, you don't have to create Syste.Object, you can simple do this:
$objectArray = #();
gci | % {
$objectArray += #{
'Key1' = 'Value 1'
'Key2' = 'Value 2'
}
}
If you need to compare complex objects, you can build them with #{} and then use Compare-Object on the two objects, just another idea.
For example, this will get a file listing of two different directories, and tell me what file exists or doesn't exist between the two directories:
$packages = (gci $boxStarterRepo -Recurse *.nuspec | Select-Object -ExpandProperty Name) -replace '.nuspec', ''
$packages += (gci $boxStarterPrivateRepo -Recurse *.nuspec | Select-Object -ExpandProperty Name) -replace '.nuspec', ''
$packages = $packages | Sort-Object
Compare-Object $packages $done
Im pulling out information from AD to output alot of information to be handled at a later point.
I need to export this to a csv or something so i can have a rollback ( in effect import file so i can handle each object like i can with the hashtable below. Especially that inside $_.MemberOf).
The hashtable im struggling with handling is:
$logging1 = #{
MemberOf="$users.MemberOf"
OriginalOU="$Ou.DistinguishedName"
DisabledWhen="$descriptionDisabled"
}
$logging = #{$users.SamAccountname = $logging1}
what the rest of the script does is(that isnt pasted ) is:
iterate through a bunch of ou's. Getting users last logged in -90 days ago or more then passing them to the HashTable in question which i like to append to a file to be imported at a later time.
I have been googling for hours without getting somewhere
here is the complete script:
import-module ActiveDirectory
$descriptionDisabled = get-date -Format yyyyMMdd
$Loggdir = "C:\temp"
$array = #{}
$loggname = get-date -f yyyyMMdd
$90days = ((get-date).AddDays(-90))
$searchBase = 'OU=someou4,OU=someou3,OU=someou2,OU=someou1,DC=name,DC=NO'
$ExclusionList = Someexlutions
$OUlist = Get-ADOrganizationalUnit -SearchBase $searchBase
foreach ($Ou in $OUlist)
{
$ExpiredADusers = get-aduser -Filter {(LastlogonDate -le $90days)
-and (Enabled -eq $True)}
-SearchBase $ou.distinguishedname
-Properties *
#Get information about every user and their groupmembership
foreach ($Users in $ExpiredADusers)
{
$users = get-aduser -Identity $users.SamAccountName
-Properties *
| Select-Object -Property SamAccountName, MemberOf,
DistinguishedName, ObjectGUID
$logging = #{$person.SamAccountName = #{
MemberOf=$person.MemberOf;
OriginalOU=$Ou.DistinguishedName;
DisabledWhen="$descriptionDisabled"}}
$Array += $Logging
}
}
$array | export-clixml -path somepath -noclobber
EDIT SOLVED
Solved problem and original script in question has been updated to handle Nested hashtables
Original problem
I think you can understand what im trying to do.
Im making a rollback file incase needed.
So the exported file need to be easy to import. when using hashtables you can use . notations and each objects under memberof is treated as an object
What the rollback needs is so i can iterate through the imported info to move the user back to its original OU placement and restore membership
I think EBGreen really hit the nail on the head here. You need an array of custom objects, and then you can just export it to a CSV like you want to. It's a really minor change in code too.
$Array = #()
$ForEach($User in $Users){
$logging1 = New-Object PSObject -Property #{
MemberOf=$user.MemberOf
OriginalOU=$Ou.DistinguishedName
DisabledWhen=$descriptionDisabled
}
$Array += $Logging1
}
Edit: Hm, so you want to be able to export and import full objects. You don't want a CSV then because you have nested arrays, and a CSV is not designed to handle that for export and import. You need XML, so, as mentioned above, you need to use Export-Clixml and Import-Clixml since XML can handle nested arrays. Just pipe the array to it once the array has all your data and you should be all set.
Edit2: The hashtables within hashtables issue... Ok, so we had it as MemberOf="$user.MemberOf" and that's the issue. It is converting it to a string, so it is expanding the entire $User variable, and tacking .MemberOf to the end of it. We don't really want to do it in this case, but if you want to access a property of an object from within doublequotes you need to put $() around it. For example if you wanted to include the user's distinguishedname as a part of human friendly output you could do something like:
Write-Output "$($Users.Name)'s distinguished name is: $($users.distinguishedname)"
Which would output something like:
TMTech's distinguished name is: CN=TMTech,OU=Awesome,OU=Administrators,DC=Digital,DC=Ghost,DC=net