Alex, July 21, 2016 - 7:23 pm UTC
I'd be curious to know the metrics behind "and these reports would run much faster if they used the shared pool." I would expect the difference between a hard parse vs soft to be fractions of a second. Obviously the system as a whole would be greatly improved by bind variables, but I'm skeptical about how much time that would save on one individual query being cached vs not. If we were talking data blocks I feel like that would make much more sense.
July 22, 2016 - 7:18 am UTC
Yes, that would be interesting to know...
If the statements are complex and there's a lot of library cache contention (which it sounds like there is) the hard parse time could become "significant". But we'd need to see the parse time stats to know that.