I have this scenario in a POC of Cassandra.
CREATE TABLE B (B-UID UUID, A-UID UUID, CREATED_AT timestamp , JSON text, PARENT_B-UID UUID, POSTALCODE text, CUSTOMER_TYPE text, START_DATE timestamp, END_DATE timestamp, SOME_PRICE int, PRIMARY KEY (B-UID));
24 k rpm of write / 2 k rpm of read. For 1 A-UID I'll have 67 B-IUD in most cases.
In the life cycle of my app I'll have to Search all the B-IUD for a specific A-IUD.
My question is: Is it better create the table with a composite primary key or create an index to A-IUD? The writing performance will be affected with a composite primary key?
I already read the docs at DataStax and the example they have written there is not too good for my case, at least in my understanding!!! :)