问题
This post from 2010 says
Query7: PartitionKey == "A" and (RowKey == "A" or RowKey == “F”)
This results in scanning the entire Partition "A".
Is this still the case for the current version of Azure?
Should I prefer two parallel queries or the syntax above when my partition contains only ~20 rows?
回答1:
I found an answer from Joe Giardino, Microsoft:
If you want several non contiguous row keys, then issuing separate but parallel individual queries of below form will perform better. When you specify both PartitionKey and RowKey the query is considered a "PointQuery" which is the most performant option.
a) PartitionKey == “MyPK” && RowKey == “FirstRK”
b) PartitionKey == “MyPK” && RowKey == “SecondRK"
来源：https://stackoverflow.com/questions/17407680/query-multiple-rowkeys-in-azure