Hello !

I've asked on many places,
but no definite answer.

We are using IOT tables as counters storage medium.
They are perfect for that purpose as they consists of only
PK + VALUE, or PK + VALUE1, VALUE2, VALUE3 up to some VALUEn n being small.
They preserve ACID principle, differently from NoSQL databases,
what is very important for our case.
Also as there is no need to store them in the two places,
i.e. in a HEAP table and its INDEX, they are saving space
and are much faster to be queried and updated
as they are always queried by PK
and consists mostly of that PK.
There is no need for TABLE ACCESS BY INDEX ROWID operation
which is slow, evan BATCHED one
when you have to fetch and/or update 10s of thousands records in one pass.
I'm telling all this because I'm always asked for the real scenario.

But when these tables grow big we need to introduce partitioning
and that was fine up to a point.
Now we also need SUBPARTITIONING for IOT tables,
but this feature is not implemented.
We handle it now someway clumsy using multicolumn partitioning,
but as we need RANGE partitions this is scrappy solution.
Especially (or at least): RANGE - HASH composite subpartitioning,
as the key always consists of:
a) some PERIOD (hour, day, week, month, ...) [ RANGE part ]
b) some ID [ HASH subpart ]
c) other fields (zones, counter types, ...)

Question:
Will there be in the foreseeing future developed SUBPARTITIONING for IOT tables?