Support secondary Indexes
Need to be able to sort on something other than the rowkey
Restricting us to only the row key is very limiting. We currently have to maintain our own secondary indexes for other columns and this is very cumbersome, it should be a feature of the platform.
We do still intend to support secondary indexes. However, no firm timeline has been established for when that capability will be made available. Based on our most current information, secondary indexes are not going to be released as a feature in 2011.
"Announced at PDC that it’s coming." - 4 years ago!
Is there any update on this? AWS is eating your lunch when it comes to table storage and honestly, we are considering moving everything to AWS just to get secondary indexes in DynamoDB. It's that big of a deal.
What is the table storage team doing? Table storage is literally the same as it was when v1 of Azure shipped 5+ years ago.
I would love to know when this is happening so we can plan accordingly. Scalable storage is a key piece of our architecture and it seems that Microsoft has abandoned it.
Tom Sawyer commented
On 2009, Dec 7, Mike Wickstrand (MS) said
"Announced at PDC that it’s coming"
Today is Oct 26, 2013. I really really really want to use Table Storage, but without secondary indexes and some LINQ support for aggregate operations it is not worth it.
Ted van der Veen commented
Or maybe implement tags that can be attached to items in the Table Storage? A bit like how tags are implemented in items in Windows Azure Caching?
My goodness, Microsoft. What are you guys thinking? This has been supposedly "in the works" for years now, and it still isn't here?
We currently have two supported options to choose from: (1) Table storage, which is basically just a glorified text file; and (2) Azure SQL, which is great, but scales like wet cement. We really need something somewhere in-between. Given any thought to supporting, say, MongoDB or CouchDB or Riak or... ? Those are currently *much* more full-featured than Azure table storage, and I can't for the life of me figure out why Azure wouldn't at least pick one to make a first-class citizen.
Rob Potter commented
Building up an index table could be implemented in application code, but I imagine that most developers have written this into their application, maintaining the indexes with every insert/delete operation. It is such a common scenario that I think it should be part of the storage platform.
I see AWS DynamoDB now supports it...
Can we get an update on when this is going to be supported?
This was announced as coming at the PDC conference in 2009. Three years later, I'm pretty sure we'll never see it. Microsoft has recently started a beta version of Hadoop on Azure (see https://www.hadooponazure.com)
My best bet now is that Hadoop will be the preferred approach and will meet whatever needs there may have been for secondary indexes in Azure Table Storage.
Once SQL Azure came out, there didn't seem to be any steam left behind further advancements to Table Storage. Microsoft missed the boat on this one, but I think this feature request has long since become irrelevant. There just wasn't much use for Table Storage in the way it was first released and it didn't get any real upgrades from that early version.
Oliver Weichhold commented
Still no news?
SoftwareMK: Mikael Koskinen commented
Aye, as many have mentioned, custom indexes are a must.
I have to say that I totally agree with people saying this is a MUST for azure table storage. Really, how is it possible to use this in any sensible way without custom indexing?
David Pfeffer commented
We are currently converting everything to use RavenDB on top of blobs for exactly this reason.
Epic fail. While Amazon DBS has had this feature for years it appears very likely it will never come to table. Table is on a path to obsolescence.
More than one year later (actually has been several years since this was requested by many, many people). Update the status?
Is there any status update since Feb 2011 on this. One year later?
I like to know: primary index is PartitionKey+RowKey, will the secondary index be: RowKey+PartitionKey or we can customize it
Its 2012. Any updates on when secondary indexes would be available. Should we plan on it being available in the near future or should we implement workarounds like storing the data multiple times with different RowKeys. As ray247 mentions not needing to make big code changes would be great and we do not want to put in unnecessary optimizations if secondary indexes are going to come by soon...
I just posted a question/suggestion to the Azure Storage forum on this secondary index feature. http://social.msdn.microsoft.com/Forums/en-US/windowsazuredata/thread/38f07a77-cbe1-4627-9c49-29b314dbfd42
Basically not needing to make big code changes for our existing app would be great when this feature comes out. So could you make it so that we could just label any existing column and make it the secondary index, this would help avoid major code changes. Thanks.
This feature is a must!!! Any news?
Oliver Weichhold commented
Any updates on this would be greatly appreciated!