MDEV-11586 UNION of FLOAT type results in erroneous precision
Fixing the asymmetry in the array field_types_merge_rules[][] which caused data loss when mixing FLOAT + BIGINT in UNIONs or hybrid functions: 1. FLOAT + INT = DOUBLE 2. FLOAT + BIGINT = FLOAT 3. INT + FLOAT = DOUBLE 4. BIGINT + FLOAT = DOUBLE Now FLOAT + BIGINT (as in #2) also produces DOUBLE, like the cases #1,#3,#4 do.
Showing
Please register or sign in to comment