Find newest information for every week

[Anonymous]

Hi all,

I have got a problem I couldn’t solve for weeks, and now I hope somebody of you is able to help me.

My datasets contains information about Testcases. Each dataset has got a Case ID, a Timestamp, a result and several other information the user can choose in the application like plan, configuration, cycle and so on.

Example

Case ID	Timestamp	Result
1234	02.01.2018	failed
5678	04.01.2018	failed
1234	05.02.2018	Blocked
2468	06.02.2018	failed
1234	20.02.2018	passed
5678	01.03.2018	passed
2468	15.03.2018	failed

Now I was asked to create a bar chart that shows for every calendar week the number of the latest results for each result. The difficulty is to look back and find for every calendar week the actual result of the highest timestamp for every Testcase.

The expression I experimented with is:

count(if(FirstSortedValue(TL_ResultStatusNeu, - Aggr( max(if(TL_TS<=TL_W_TSWeekID,TL_TS)),TL_CaseID,TL_W_TSWeekID))='passed',TL_CaseID))

but it doesn’t work.

I hope someone can help me.

Best regards

Michael

[sunny_talwar]

What is the expected output here?

[Digvijay_Singh]

May be this -

data used as below -

Case ID	Timestamp	Result
1234	04.01.2018	Blocked
1234	03.01.2018	failed
5678	04.01.2018	Blocked
1234	05.02.2018	Blocked
2468	06.02.2018	failed
1234	20.02.2018	passed
5678	01.03.2018	passed
5678	04.03.2018	failed
9999	05.03.2018	failed
2468	15.03.2018	failed

[Anonymous]

I need the following output:

					Counted Case IDs
Calendar Week	Failed	Blocked	Passed		Failed	Blocked	Passed
1-18	2	0	0		1234;5678
2-18	2	0	0		1234;5678
3-18	2	0	0		1234;5678
4-18	2	0	0		1234;5678
5-18	2	0	0		1234;5678
6-18	2	1	0		5678;2468	1234
7-18	2	1	0		5678;2468	1234
8-18	2	0	1		5678;2468		1234
9-18	1	0	2		2468		1234;5678
10-18	1	0	2		2468		1234;5678
11-18	1	0	2		2468		1234;5678

For every Calendar Week I need the last result set in the past. In the weeks 2 to 5 there are no new results, therefore the old one are still relevant. My problem is to find an expression that "can look" in the past over the limits of the actual Calendar Week. Furthermore I don't know, what the user will select like plan, configuration and so on (I did not mentioned it here to keep it as simple as possible) . So it can be, that not every dataset will be in the selection of the user, so that the result will be different.