Link to home
Start Free TrialLog in
Avatar of ravi doshi
ravi doshi

asked on

How to combine two CSV files with different column headers in powershell

I am looking to combine two CSV files that have different column headers in powershell.  Some are the same but others are not.  I need all data from each sheet to fall into the correct columns.  No data should be excluded.  All columns should remain after the merge.

CSV1:
User generated image
CSV2:
User generated image
What I need for final output:
User generated image
Avatar of oBdA
oBdA

Assumes that 'Source Key' is the common key, and that it is unique per table.
Will write warnings (and list them in an additional column 'Conflicts') if there's a matching 'Source Key', but one of the common fields differs.
$csvLeft =	'C:\Temp\users1.csv'
$csvRight =	'C:\Temp\users2.csv'
$csvOut =	'C:\Temp\users_out.csv'

$joinOn = 'Source Key'
$commonProps =	@('First Name', 'Last Name', 'Work Facility Location', 'Work Address 1', 'Work City', 'Work State', 'Work Zip', 'Work Country')
$allProps =		@('Source Key') + $commonProps + @('Work Floor', 'Work Building', 'Device 1 Type', 'Device 1 Address', 'Conflicts')
$lookup = @{}
Import-Csv -Path $csvRight | ForEach-Object {$lookup[$_.$joinOn] = $_}
$(
	Import-Csv -Path $csvLeft | ForEach-Object {
		$conflicts = @()
		If ($lookup.ContainsKey($_.$joinOn)) {
			ForEach ($prop in $commonProps) {
				If ($_.$prop -ne $lookup[$_.$joinOn].$prop) {
					$conflicts += $prop
					Write-Warning "Key '$($_.$joinOn)': conflict in column '$($prop)'!"
				}
			}
			If ($conflicts) {
				$_, $lookup[$_.$joinOn] | Select-Object -Property *, @{n='Conflicts'; e={$conflicts -join '; '}}
			} Else {
				$_ | Select-Object -Property *, @{n='Device 1 Type'; e={$lookup[$_.$joinOn].'Device 1 Type'}}, @{n='Device 1 Address'; e={$lookup[$_.$joinOn].'Device 1 Address'}}
			}
			$lookup.Remove($_.$joinOn)
		} Else {
			$_
		}
	}
	$lookup.Values
) | Select-Object -Property $allProps | Export-Csv -NoTypeInformation -Path $csvOut

Open in new window

Avatar of ravi doshi

ASKER

This worked but could you please help me understand what lines 22 & 23 accomplish?  The screenshots I included were just a few of the columns so I added the rest of the fields to the common properties and all properties sections and was expecting to need to do the same below but I didn't have to modify line 23.  Why?
Here in the final code that worked for me:

$csvLeft =	'D:\RD_Test\left.csv'
$csvRight =	'D:\RD_Test\right.csv'
$csvOut =	'D:\RD_Test\final.csv'

$joinOn = 'Source Key'
$commonProps =	@('First Name', 'Last Name', 'Employee ID', 'Job Title', 'Company', 'Time Zone', 'Division', 'Language', 'Work Facility Location', 'Work Address 1', 'Work City', 'Work State', 'Work Zip', 'Work Country', 'Device 1 Type', 'Device 1 Address', 'Device 1 Description', 'Group 1', 'Group 2')
$allProps =		@('Source Key') + $commonProps + @('Login ID', 'Work Floor', 'Work Building', 'Device 2 Type', 'Device 2 Address', 'Device 2 Description', 'Device 4 Type', 'Device 4 Address', 'Device 4 Description', 'Conflicts')
$lookup = @{}
Import-Csv -Path $csvRight | ForEach-Object {$lookup[$_.$joinOn] = $_}
$(
	Import-Csv -Path $csvLeft | ForEach-Object {
		$conflicts = @()
		If ($lookup.ContainsKey($_.$joinOn)) {
			ForEach ($prop in $commonProps) {
				If ($_.$prop -ne $lookup[$_.$joinOn].$prop) {
					$conflicts += $prop
					Write-Warning "Key '$($_.$joinOn)': conflict in column '$($prop)'!"
				}
			}
			If ($conflicts) {
				$_, $lookup[$_.$joinOn] | Select-Object -Property *, @{n='Conflicts'; e={$conflicts -join '; '}}
			} Else {
				$_ | Select-Object -Property *, @{n='Device 1 Type'; e={$lookup[$_.$joinOn].'Device 1 Type'}}, @{n='Device 1 Address'; e={$lookup[$_.$joinOn].'Device 1 Address'}}
			}
			$lookup.Remove($_.$joinOn)
		} Else {
			$_
		}
	}
	$lookup.Values
) | Select-Object -Property $allProps | Export-Csv -NoTypeInformation -Path $csvOut

Open in new window

That can't really have produced the correct results, unless you have no "Source Key" that is identical in both tables, that is, you have no rows that can be merged.
Lines 21 and 23 handle the case where 'Source Key' is the same in both tables; if you don't have that situation at all, that is, if you have two totally independent tables that just differ in a few columns, the script can be a lot easier.
21 handles the case where "Source Key" is the same in both tables, but one of the common fields differs. In this case, both rows will be returned, with the conflicting fields added.
23 handles the case where "Source Key" is the same in both tables, and the rows can be joined by adding the contents of the columns that are only in the "right" table to the current "left" row (that's what the calculated properties - @{n='...'; e={}} - do).
23 requires one calculated property for each "right" column that is missing in the "left" table.
Thanks for the explanation.  No source key is the same in either file.  One file is pulled from our HR system and the other is pulled from our Active Directory.  There is no overlap in either file - completely different set of users.  The only issue was that there are different columns in each file.  Please post the simpler version, just for my knowledge, if it's not too much trouble.
ASKER CERTIFIED SOLUTION
Avatar of oBdA
oBdA

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Thanks for all the explanations.  I don't like to just copy code.  Understanding how it works is important to me!