The Vertica Forum recently got a makeover! Let us know what you think by filling out this short survey.
Multi-customer schema design
Hi , As part of our multi-talented schema design (for SASS project) we check the option to store data of different customers on the same column tables , the data is mainly related to performance metrics , so one of our tables will looks like : Table ======== Customer_id , Key1 , key2 ,Key3 , Metric1, Metric2,Metric2 Query ======== Select Metric1, Metric2 From Table Where Customer_id =123 And Key1=’A’ and key2 =’B’ and Key3=’C’ Assuming the query’s we run do not filter the Metrics Values only the Key’s (As above) , I have a performance concern regarding column reconstruct extra I/O, assuming vertica implement late materialize , vertica will have the read the entire column data of Metric1, Metric2 to bring the positions columns and then be merge with potions of Customer_id, Key1 , key2 ,Key3 .So in fact we will read matric's data of diffrent customers , If this true fact , we may consider to build different multi-talented schema . Please advice