Powershell：尝试组合两个不使用相同列的数组

Question

I have two arrays imported from csv files, which from here on i will refer to as the master array and update array.

The master array has three extra columns on the front, and three extra columnns on the back.

Each day i get a new update array that i need to do two things with.

A) Remove any rows on the master that do not appear in the update

B) Add any rows that appear in the update but not the master to the master

I am still fairly new to powershell, and scripting in general(mostly self taught) and can't figure out how to approach this. I know there's a compare-object command, so I can get a list of which rows match pretty easily, but I'm not sure how to combine them the way I want.

Edit: The master array entries have this information:

ef: true
ea: true
rem: true
accountname: example1
Enabled: True
PasswordLastSet: 01/08/2002 13:14:19
whencreated: 01/08/2002 13:14:19
Description:
Owner Email: johnsmith@email.com
givenname: John
sn: Smith
manager: John Doe
Level2: Person 1
Level3: Person 2
Level4: Person 3

While the updates only have:

accountname: example1
Enabled: True
PasswordLastSet: 01/08/2002 13:14:19
whencreated: 01/08/2002 13:14:19
Description:
Owner Email: johnsmith@email.com
givenname: John
sn: Smith
manager: John Doe

Answer 1

Assuming the accountname column can be used as the unique key that ties the two arrays together, you could use something like the below script. It creates a third array and then overwrites the master array csv once completed.

$arrmaster = import-csv c:\temp\arrmaster.csv
$arrupdate = import-csv c:\temp\arrupdate.csv
$arrworking = @()

foreach ($rowupdate in $arrupdate){
    $rowmaster = @($arrmaster | where {$_.accountname -eq $rowupdate.accountname})
    if ($rowmaster.Count -lt 1){
        Write-Debug "Could not find record for $($row.accountname)"
    } 
    if ($rowmaster.Count -gt 1){
        Write-Debug "Found duplicate records for $($row.accountname)"
    }
    if ($rowmaster.Count -eq 1){
        $rowworking = "" | select ef,ea,rem,accountname,Enabled,PasswordLastSet,whencreated,Description,"Owner Email",givenname,sn,manager,Level2,Level3,Level4
        $rowworking.ef              = $rowmaster.ef
        $rowworking.ea              = $rowmaster.ea
        $rowworking.rem             = $rowmaster.rem
        $rowworking.accountname     = $rowupdate.accountname
        $rowworking.Enabled         = $rowupdate.Enabled
        $rowworking.PasswordLastSet = $rowupdate.PasswordLastSet
        $rowworking.whencreated     = $rowupdate.whencreated
        $rowworking.Description     = $rowupdate.Description
        $rowworking."Owner Email"   = $rowupdate."Owner Email"
        $rowworking.givenname       = $rowupdate.givenname
        $rowworking.sn              = $rowupdate.sn
        $rowworking.manager         = $rowupdate.manager
        $rowworking.Level2          = $rowmaster.Level2
        $rowworking.Level3          = $rowmaster.Level3
        $rowworking.Level4          = $rowmaster.Level4

        $arrworking += $rowworking
    }
}

$arrworking | Export-Csv -Force -NoTypeInformation c:\temp\arrmaster.csv

Answer 2

Not tested, but I think this should work:

$MasterFile = 'c:\somedir\master.csv'
$UpdateFile = 'c:\somedir\update.csv'

$master= @{}
$update = @{}

import-csv $MasterFile |
ForEach-Object { $master[$_.accountname] = $_ }

import-csv $update |
ForEach-Object { $update[$_.accountname] = $_ }

#Get Master entries contained in Update 
[array]$NewArray = $master.keys |
 Where-Object { $update.keys -contains $_ } |
 ForEach-Object { $master[$_] }

 #Get Updates not in Master
 $NewArray += $update.keys | 
 Where-Object { $master.keys -notcontains $_ } |
 ForEach-Object { $update[$_] }

 $NewArray | Export-Csv 'c:\somedir\new_master.csv' -NoTypeInformation

That starts by loading each of your arrays into a hash table, indexed by the accountname. Then the keys are used to extract the master entries that have an accountname that appears in the update keys and load that into a new array. Then the process is reversed and the update keys compared to the master keys, and any entries that do not have a matching key in the master are added to the array. Then the array is exported to csv.

The CSV export will create it's header row from the first entry, and add the necessary commas for any objects in the array afterward that are missing properties. You don't have to worry about adding the missing properties to the update entries as long as they're added after the master entries.

Answer 3

Ok, again based off the assumption that AccountName is a unique identifier that both lists would have in common you can run this:

$Master = Import-CSV Master.csv
$Update = Import-CSV Update.csv
$T2Keys = $Master|gm|?{$_.MemberType -match "Property"}|Select -ExpandProperty Name
$T1Keys = $Update|gm|?{$_.MemberType -match "Property"}|Select -ExpandProperty Name
$KeysToAdd = $T2Keys|?{$T1Keys -notcontains $_}
$NewMaster = @()
$NewMaster += $Update | ?{!($Master.accountname -contains $_.accountname)}
$KeysToAdd|%{$NewMaster|Add-Member $_ ""}
$NewMaster += $Master | ?{$Update.accountname -contains $_.accountname}
$Newmaster| Select ef,ea,rem,accountname,enabled,passwordlastset,whencreated,description,'owner email',givenname,sn,manager,level2,level3,level4|Export-CSV NewMaster.csv -notype

Ok, that will import a CSV for the master list and the updates list. If you already have those as objects then skip the import-csv lines. Then it gets all properties from both, and figures out which ones to add to the updates (the 6 that the master has that the updates doesn't). It then creates an empty array and adds all records from the Updates list to it that aren't in the master list. Then it adds the missing fields, and adds all the records from the master list that are in the updates list. Then it exports it to a CSV. So it does what you asked:

• Gets all records from the master list that are in both lists.
• Adds records from the update list that are missing from the master list.

Edit: The reason I had asked if you had searched is that 95% of my answer there was almost copied and pasted from this question that I answered just under a month ago. But hey, it's all good, not that hard for me to copy and paste to get you an answer, and I kinda knew what I was looking for anyway. I don't know that the other question's title would have been indicative that it had what you needed.