Pass/Fail counter question. [Help]

[altenbach]

Again, turn the "3" into a control if a run has a settable number of tests, or you can calculate it automatically from the number of unique test or the number of elements having the same sample name, for example.

[tbob]

---

@altenbach wrote:

How do you tell the difference betweeen:

• One sample failing all three tests
• Three different samples failing one test each? 

---

In the testing world that I live in, if a sample fails any test or number of tests, it is a failed unit.  If three samples fail one test each, then I have three failed units.  We usually store measurements, limits, pass/fail flag, and such in a database.  We have tools to extract data from the database to do in depth analysis.

[tbob]

I had come up with this code but didn't post it because Jim beat me to the punch, and his code seemed more elegent.  There are so many ways to skin a cat (knife, razor blade, hatchet, axe, hydrocholric acid, etc...)

[altenbach]

Sorry, I wasn't clear. I was wondering how your program can tell the difference. 😉

(Sorry, I don't have the "oktk" installed on the current rig, so I cannot test...)

[jcarmody]

Here's a new approach.  It doesn't assume anything about the structure of the array.  The only assumptions are that each TSample will have the same name in each test and that a pass will be exactly "Pass".  The OpenG array tools make this much easier.  🙂

[tbob]

---

@altenbach wrote:

Sorry, I wasn't clear. I was wondering how your program can tell the difference. 😉

---

It can't.  But it doesn't matter.  One test failure or three test failures is still a failed unit.

[altenbach]

---

@tbob wrote:

---

@altenbach wrote:

Sorry, I wasn't clear. I was wondering how your program can tell the difference. 😉

---

It can't.  But it doesn't matter.  One test failure or three test failures is still a failed unit.

---

Yes, but three test failures could be anywhere from 1-3 failed units. Your new code is better. 😉

[tbob]

Altenbach, I think you were looking at Jim's code that I had just added some more code at the end.  Yes this code is wrong.  I modified it without thingking.  The code I last published is my code.  It accurately gives the correct number of passed samples.

I think we got crossed up with the latency in posting.  I posted at a time when you had already posted and I didn't see it, and vice versa.

[altenbach]

Yes, one of these race conditions ....

My code is very similar to yours, just maybe a little tighter. 😉

[jcarmody]

---

@tbob wrote:

[...] It accurately gives the correct number of passed samples. [...}

---

Unless the rows per sample is not uniform, or if the array isn't properly sorted.