In Organization, we are palnning to move to teradata enterprise datawarehouse. But all the datawarehouse products will work fine in the normal scenorios, how teradata behaves in the following scenorio,
We have dimension which is having 10 M records and a helper table with 20 M records and fact with the 20 M records. When the reports hits the helper tables the it dies.
The tables are properly indexed with the help of the DBA's. How does the Teradata behaves in this scenorio.
There's almost no difference between an Integer PI and a 100 byte multi-column VarChar PI, because Teradata uses hashing. Of course there's more overhead to calculate a hash value from 100 byte compared to 4 bytes, but this is negligible.
For joins it's similar: first both hash values are compared (each 4 bytes just like an integer) and only if those are the same, then the columns are compared.