How to populate field based on groupings? - sql

Question

I am trying to create a PowerShell (5.1) script (Open to SQL suggestions as well using SSMS with SQL Server 2014) to eliminate manual effort of updating a very large data file exported in CSV format.
Here is what the raw data set looks like that needs updated:
Parent ID | Parent Owner | Sub ID | Sub Owner | Sub Hours
A aA Rob Green 0
A aB Rob Green 6
B aA Jane Doe 4
B aB Jane Doe 10
B aC Bob Smith 18
C cA Jane Doe 0
C cB Jane Doe 6
D dA Bob Smith 0
D dB Bob Smith 6
E dE Joe Brown 0
As you can see, Parent IDs can have one or many Sub Owners and Sub IDs.
My goal is to populate the Parent Owner field based on the following criteria:
For every Parent ID set
If there is only one distinct Sub Owner, then that Sub Owner should be the Parent Owner for all corresponding Parent IDs.
If there is only one occurrence of a Parent ID then that Sub Owner should be the Parent Owner for that Parent ID.
If there are multiple Sub Owners for the Parent ID, the Sub Owner with the highest summed Sub Hours should be the Parent Owner for every occurrence of that Parent ID.
To clarify, the criteria applies to the raw data above like so:
Parent ID "A" applies to criteria 1
Parent ID "B" applies to criteria 3
Parent ID "C" applies to criteria 1
Parent ID "D" applies to criteria 1
Parent ID "E" applies to criteria 2
This is what I expect the data above to look like after completed:
Parent ID | Parent Owner | Sub ID | Sub Owner | Sub Hours
A Rob Green aA Rob Green 0
A Rob Green aB Rob Green 6
B Bob Smith aA Jane Doe 4
B Bob Smith aB Jane Doe 10
B Bob Smith aC Bob Smith 18
C Jane Doe cA Jane Doe 0
C Jane Doe cB Jane Doe 6
D Bob Smith dA Bob Smith 0
D Bob Smith dB Bob Smith 6
E Joe Brown dE Joe Brown 0
My biggest struggle is criteria 3. I cannot wrap my head around how to do this. Can anyone give me an idea of how I can get my expected output using PS or SQL?
Any help would be GREATLY appreciated!

Answer 1

I vowed myself that SQL is strictly taboo for me. However, below's an example of pure PowerShell solution (and I'm pretty sure that it's convertible to SQL simply):
# mimic SQL output
$SqlOutput = #"
Parent ID|Parent Owner|Sub ID|Sub Owner|Sub Hours
A||aA|Rob Green|0
A||aB|Rob Green|6
B||aA|Jane Doe|4
B||aB|Jane Doe|10
B||aC|Bob Smith|18
C||cA|Jane Doe|0
C||cB|Jane Doe|6
D||dA|Bob Smith|0
D||dB|Bob Smith|6
E||dE|Joe Brown|0
"# | ConvertFrom-Csv -Delimiter '|'
# compute an auxiliary variable
$SqlOutputGroups = $SqlOutput |
Group-Object -Property 'Parent ID', 'Sub Owner' |
ForEach-Object {
New-Object psobject -Property #{
'Parent ID' = ( $_.Name -split ', ')[0]
'Sub Owner' = ( $_.Name -split ', ')[1]
Hours = ( $_.Group |
Measure-Object -Property 'Sub Hours' -Sum).Sum
}
}
# compute Criterium3 as a hashtable
$SqlOutputCriterium3 = #{}
$SqlOutputGroups | Group-Object -Property 'Parent ID' |
ForEach-Object {
$SqlOutputCriterium3[$_.Name] = ($_.Group |
Sort-Object -Property Hours |
Select-Object -Last 1).'Sub Owner'
}
# apply Criterium3
$SqlOutput | ForEach-Object {
$_.'Parent Owner' = $SqlOutputCriterium3.$($_.'Parent ID')
}
# show result in a table format
$SqlOutput | Format-Table -AutoSize
Output: D:\PShell\SO\45963820.ps1
Parent ID Parent Owner Sub ID Sub Owner Sub Hours
--------- ------------ ------ --------- ---------
A Rob Green aA Rob Green 0
A Rob Green aB Rob Green 6
B Bob Smith aA Jane Doe 4
B Bob Smith aB Jane Doe 10
B Bob Smith aC Bob Smith 18
C Jane Doe cA Jane Doe 0
C Jane Doe cB Jane Doe 6
D Bob Smith dA Bob Smith 0
D Bob Smith dB Bob Smith 6
E Joe Brown dE Joe Brown 0
Note that Criterium 3 covers criteria 1 and 2 but does not suffice if more Sub Owners have the same highest sum of Sub Hours for a particular Parent ID (e.g in case of B||aA|Jane Doe|8 instead of B||aA|Jane Doe|4 in the above example, then Jane Doe has sum of Sub Hours =18 as well as Bob Smith in Parent ID=B).