Studying your data
As you already know by now, tables have one single index, which is the combination of both keys. This means that you are unable to add secondary indexes, as you can in the case of relational databases. So, to achieve optimal scalability and query efficiency, you need to analyze your data by determining your partition sizes and understanding how you will query your data.
Partition sizing
More partitions mean better scalability and load balancing. Partition size is the number of entities in one single partition. Understanding your data and choosing your partition key will determine your partition count and size. You might design your table to have only one partition key and on, the other extreme, you might choose the finest level of granularity and insert a unique key for each entity:
Choosing a single value for the PartitionKey: With a low entity count you can perform batch transactions on the entire table. But with a larger number of entities, scaling will be limited and...
