This is a simplified table of Cancer-examinations.:

```
Examination Round Corrected
```

```
exam001 round001 0
exam002 round001 1
exam003 round002 0
exam004 round002 1
exam005 round003 0
exam006 round003 0
```

I want to write a query that removes exam001 and exam003.

Lets look at round0001. It has exam001 and exam002 connected to it.

But since exam002 is corrected, we know that there was something wrong

with exam001. Hence, I want to remove it.

I do not want to delete any exams for round003, since none of them

are corrected.

So in pseudocode:

foreach round

delete all exams where corrected == 0

if round has exam where corrected == 1

It is given that every round has exactly 2 exams.

How would I do this?