Toad World® Forums

Health check type comparisons for joins

Hi,

toad health check shows for the check of type matching of join columns number vs number(8) as an issue.

For my part i dont see any reason for this as there is no physikal difference in those two and i wonder if there will ever be an implicit type conversion in those cases.

Any justification for the check marking this “mismatch” as an potential performance impact that i am not aware?

Thanks in advance!

Regards chris

if the NUMBER column contains data that wouldn’t fit in the NUMBER(8) column, there could be a problem with your join. If that will never happen, then it would be best to define the columns the same way.

Thanks for your answer.

However the issue is still unclear for me.

What could be the problem if the is no difference in the way number and number(8) is stored physically?

Furthermore you will never see something like an implicit type conversion in the execution plan.

Is there hard fact like entries in the trace file e.g., that prove that “there could be a problem”?

Regards

Chris

I think the only potential problem between NUMBER and NUMBER(8) is that the NUMBER column can store some data that the NUMBER(8) column cannot.