Help developing a Qlik expression for transaction communication data

[mb_me]

Hey everyone, I'm at wits end with this problem. Any help is greatly appreciated!

Scenario:
- Determine average response time from customer -> support agent, and average response time from support agent -> customer, for each case id in the dataset. 

Dataset example:

Case id	case creation date	transaction type code	transaction name	transaction date	transaction timestamp (hh:mm:ss)
1	11/15/2024	1	Customer submission	11/15/2024	06:32:30
1	11/15/2024	5	support provides solution	11/15/2024	07:15:20
1	11/15/2024	2	customer denies solution	11/15/2024	07:25:15
1	11/15/2024	5	support provides solution	11/15/2024	07:35:10
1	11/15/2024	2	customer denies solution	11/15/2024	08:15:18
1	11/15/2024	5	support provides solution	11/15/2024	08:32:53
1	11/15/2024	3	customer accepts solution	11/15/2024	08:45:32
1	11/15/2024	8	support closes case	11/15/2024	09:18:36

So support response times would look at the time difference between message types 5 or 8 (support replies), and 1, 2, or 3 (customer questions). 
Customer response times would look at the time difference between message types 2 or 3 (customer replies), and 5 (support answers).

Has anyone had experience answering a similar problem in their work?

Note: I don't have developer rights for the app that provides this dataset, so I am stuck using the Qlik expression editor.

[Qrishna]

Usually we need a sorting field for this kind of problems, as you said you dont have access to load editor, you can try below, but make sure the order is always that shown in th table above (Ascending Time sort like above):

Interval(Aggr(if(RowNo(total) = 1
, 0
, if(RowNo(total) <> 1 and [Case id] = above([Case id]) and Subfield(Capitalize([transaction name]), ' ', 1) <> Above(Subfield(Capitalize([transaction name]), ' ', 1))
, [transaction timestamp (hh:mm:ss)] - Above([transaction timestamp (hh:mm:ss)])
, 0
)
)
, [transaction timestamp (hh:mm:ss)])
, 'hh:mm:ss')