If we want to support the expr of key, the different load plan will be need
between base and rollup.
The key of rollup needs to be transformed by the to_mouth function before
the storage engine aggregate the value column.
>From the perspective of demand, users only need to be able to store day,
mo
If expressions are supported, the storage engine will changed when the
function column is built automatically.
Also the mv selector need a expression calculation tool to judge whether mv
could be selected for query.
After Doris support the count_distinct and count, I will consider this
feature.
陈
That could be good.
However, if expressions are supported, the function of automatically selecting
rollup may fail.
Because the reason why Doris can automatically select the rollup is based on
keeping the column values consistent.
When the column value generates a new value through the expressi
Hi Ling
Very glad to see this proposal.
We can discuss whether to support expression operations.
For the case that expressions are applied on the KEY column. For example we
can easily get a monthly table
through a rollup which is defined by "group by to_month(day_column)".
For the case that exp
Hi everyone,
*The status of materialized view 1.0*
In the present, we have supported the materialized views in Doris 0.12
version. The materialized view selector supports to select the most
efficient mv and rewrite the SQL to query against the selected mv instead
of the base table.
For query resul
