Alex, It seems that the described problem is not a problem at all. The mentioned reason (is not obvious what is a cache name and what is next hierarchical element) will exist forever. If we will replace dots by underscores (or escaped dots, or something else) still will exist the possibility of naming cache in such a way where your reason will be actual.
On Wed, Sep 16, 2020 at 11:04 AM Alex Kozhenkov <[email protected]> wrote: > > Ok, then can we replace dots in cache names only in metric names, and maybe > in ignite 3.0 restrict all unwanted symbols directly to cache names? > > вт, 15 сент. 2020 г. в 16:41, Nikolay Izhikov <[email protected]>: > > > Hello, Alex. > > > > We can’t restrict dot in the cache name because it will break backward > > compatibility of Ignite. > > > > > 15 сент. 2020 г., в 16:38, Alex Kozhenkov <[email protected]> > > написал(а): > > > > > > Yeah, that would be even better for all cases. > > > > > > вт, 15 сент. 2020 г. в 15:51, Ilya Kasnacheev <[email protected] > > >: > > > > > >> Hello! > > >> > > >> This seems suspiciously similar to > > >> https://issues.apache.org/jira/browse/IGNITE-7264 > > >> > > >> Maybe we need to restrict cache names after all. > > >> > > >> Regards, > > >> -- > > >> Ilya Kasnacheev > > >> > > >> > > >> вт, 15 сент. 2020 г. в 13:49, Alex Kozhenkov <[email protected]>: > > >> > > >>> Hi, guys! > > >>> > > >>> Cache names may contain dots, but in metic names we create hierarchy > > with > > >>> dots. > > >>> So with the cache name "Person.statistic" we will have a metric > > >>> "io.statistics.sortedIndexes.SQL_PUBLIC_Person.statistic._key_PK" which > > >> is > > >>> not obvious what is a cache name and what is next hierarchical element. > > >>> > > >>> I suggest replacing dots with the special symbol during metrics names > > >>> construction. > > >>> > > >> > > > >
