Clarification on synthetic partition key usage

Question

Hello Team,

I have a read heavy container (Total number of records ~300M,Distinct records for A ~20M,
Distinct records for B ~20K).A is chosen as the partition key. A large number of records with "" value for A is expected. But reads based on A="" are not expected. To solve the hot partition for "" value thinking of creating
a synthetic partition key by concatenating values of A and B.

Following are the main queries against the container.

Select from c where c.A=sample or c.B=sample
Select from c where c.A in (comma seperated values)

Please help me to understand if the synthetic partition key helps the above queries. How will the query identify the partition to be used? How will it affect the query performance?

Answer 1

Hi @Litty Kuriakose , welcmoe to Microsoft Q&A forum.

Using synthetic partition key seems like the right idea in this specific situation, however the queries mentioned above will not use the synthetic partition key because synthetic keys are created as concatenation of properties A and B (lets say "AB"), until we use "AB" in the queries it will not be used. As an example, below query will use the key:

Select from c where c.AB = sample  

Please let us know if this helps or we can continue the discussion.

Answer 2

Certamente a produção do seu padrão de publicidade é em VCards.