ok

due to buisness reasons, i can not post ddl and sql but i can try to
give example close to real situation

table A     |   table B     |   table C
aa int,      |    ba int,         |    ca int,
ab int,      |    bb int,         |    cb int,
ac int,      |     bc date        |  cc int,
ad date,   |              |  cd date
ae date,   |              |
af date,    |               |
ag date    |              |

there is mdc on each table on column ad date, bc date, cd date
there is regular index like

index a on table a
aa asc, ab asc, ac asc

index b on table b
ba asc, bb asc

index c on table c
ca asc, cb asc, cc asc

and sql is like

select (some coulmns from table a,b,c) from a,b,c
where a.aa = b.ba and a.ab = b.bb and a.aa = c.ca and a.ab = c.cb
and ag between (current date - 42 months) and (current date)
and bc between (current date - 42 months) and (current date)

now it is using dimansional block index because its cluster ratio is
100% but skipping index a or b or c which has column used in join

those indexes has bad cluster ratio which can not avoided

how should i go about this
i think sql will improve if it use regular indexes along with block
indexes