Hi Naga,
first thing is to check the statistics. And information about 2nd table is, indeed, very needed.
Apart from that, if you select only a few columns(4) from a table, a distribution might be a good choice, actually. Because redistribution takes place only for the columns that you select.
How do you know that it is causing a bottleneck? Is there a skew when the table gets redistributed?
p.s. also, you might want to introduce IDs instead of varchars for PI column, if possible. But that does not relate to the redistribution, though.
Regards,
Vlad.
Hi Naga,
first thing is to check the statistics. And information about 2nd table is, indeed, very needed.
Apart from that, if you select only a few columns(4) from a table, a distribution might be a good choice, actually. Because redistribution takes place only for the columns that you select.
How do you know that it is causing a bottleneck? Is there a skew when the table gets redistributed?
p.s. also, you might want to introduce IDs instead of varchars for PI column, if possible. But that does not relate to the redistribution, though.
Regards,
Vlad.