Reputation: 973
I am trying to implement a structure like the following: with property name "u:uid ", so that I can do a query like [(u:112 eq 1) or (u:118 eq 1) or (u:119 eq 1)] . It can potentially get the work done, just that I fear it would not be good for situation like searching for hundreds of uid. And when the application grows the number of "columns" can grow to tens of thousand. Can anyone familiar with the azure table architecture shed some light on this? And would there be alternative way to implement the same structure?
Upvotes: 0
Views: 111
Reputation: 136136
So a few things here:
:
in it.u
is an attribute name in the table, querying just on it would result in full table scan as your query does not include PartitionKey
. While this may not be an issue if your table is small however it would create a lot of problems as your table grow big. If the uid
is unique, I suggest you set its value as the PartitionKey and then you could query like
(PartitionKey eq '112') or (PartitionKey eq '118') or (PartitionKey eq '119')
I would strongly recommend reading this blog post from Windows Azure Storage team before proceeding: http://blogs.msdn.com/b/windowsazurestorage/archive/2010/11/06/how-to-get-most-out-of-windows-azure-tables.aspx.
Upvotes: 1