Performance impact upon increasing number of columns
I would like to know what is the performace impact upon increasing number of columns.
We are having a Vertica cluster (6 nodes) with two big fact tacles (360 column each) and they partition by date_id (365 days). we need to increase the number of column on each of the fact tables by arroud 200 columns.
1. we know that teoreticly Vertica can support maximum of 1600 columns per tabel.
2. Is the insertion performance (time) will dramaticly fail? or it will decrease by the relative number of columns will be added?
3. It is possibly to say that the performance drop off (all asspect including tupple mover operations) compared to increasing number of columns
4. Is their a performance breaking point upon increasing number of columns?