Is it possible to use the DISTINCT on specific columns? For instance: a table has ColA,ColB,ColC and I want my SELECT DISTINCT to focus on ColA and
ColB but also return ColC. However, ColC is not part of the DISTINCT clause.
Re: Optimizer Issues with 10.00.xC8 ... That would sidestep the low selectivity of the colC index.... We have some queries that search based on ColA, some that search based on ColA and ColC combined, and some that search on ColC alone. ... I, and several other customers, complained about the problem and one of the big advances in the optimizer in 7.31/9.30 was for the engine to delay the final calculations of cost for these queries until it had the replacement values at cursor open/execute time. ... this _only_ appears to be the case if there are distributions on the table. ... (comp.databases.informix)
Re: Optimizer Issues with 10.00.xC8 ...query runs), but the point is that in IDS 10.00.FC6, these queries ran ... one is a composite, unique index on colA, colB ... and the other is a nonunique index on colC... If I run full statistics with distributions... (comp.databases.informix)
Re: Optimizer Issues with 10.00.xC8 ... That would sidestep the low selectivity of the colC index.... We have some queries that search based on ColA, some that search based on ColA and ColC combined, and some that search on ColC alone. ... So if it gets a ColC=0 request first, it picks the ColA index, and does so for subsequent executions as well. ... (comp.databases.informix)
Optimizer Issues with 10.00.xC8 ... Has anyone else been having optimizer issues with IDS 10.00.xC8? ... one is a composite, unique index on colA, colB and the other is a nonunique index on colC. ... If I run full statistics with distributions, the optimizer picks the index on colC every time, which is Very Bad, and takes forever. ... (comp.databases.informix)
