CBO uses different array expression for histogram generation and selectivity calculation

For predicate: any i in uarr_i SATISFIES i > 34741 END and norm_i <= 60335
For index key: DISTINCT ARRAY `i` FOR i in `uarr_i` END

(distinct (array _usv_1 for _usv_1 in uarr_i end) is used to generate histogram.

But in selectivity calculation (distinct (`uarr_i`)) is used to look up for histogram, and histogram returns nil.

update statistics for ks1(DISTINCT ARRAY `i` FOR i in `uarr_i` END) and
update statistics for ks(DISTINCT uarr_i)
may need to identify each other and switch when necessary.