Azure Search
Azure Search is a search-as-a-service solution that allows developers to incorporate a sophisticated search experience into web and mobile applications without having to worry about the complexities of full-text search and without having to deploy, maintain or manage any infrastructure
-
Modelling Complex Types in Indexes
When modelling an index the data types are restrictive. There are simple types and collections. There is nothing that allows us to model complex types e.g.
...
"name": {
"first": "Ericka",
"last": "Banks"
},......The oData spec allows for complex types
http://www.odata.org/documentation/odata-version-2-0/overview/
963 votesWe are ready to onboard more customers who are interested in participating in the private preview of Complex Types. If you’re interested, please contact us at azscustquestions@microsoft.com.
Please note that the private preview of Complex Types is only accessible via the REST API. We are working on .NET SDK support for the Public Preview.
Also, if you get an error message when you email azscustquestions, don’t worry — we are receiving your requests and will reply to you individually.
Thanks!
-
Allow to upgrade the pricing tier
Everywhere in the portal, we can increase/decrease the pricing tier based on our needs.
With the introduction of Azure Search Basic, we need a way to upgrade to the next level (Standard) when we are close to reach the service limits.The only way today to go around this is to bring the index down and create another one.
844 votes -
Provide a local emulator or increase the number of indexes on the free tier for development use
It sucks that if I have more than 3 indexes I have to create a paid tier for development. We do not need performance or size. Just more index to simulate the production environment which contains more indexes.
704 votes -
Support schema changes with automated reindexing
Azure Search should support safe schema changes, such as deletion of fields, or maybe even safe field type changes (e.g., int -> string, or string -> string collection)
555 votes -
Add aggregations functionality
Expose functionality equivalent to Elasticsearch aggregations
439 votes -
Support case-insensitive comparison in filters
Searching with text=<xyz> performs a case-insensitive search. However, testing string fields in filters is case-sensitive. There should be a way to do a case-insensitive string comparison in filter expressions.
423 votes -
Support "swapping" indexes
Let's say you want to make some changes to an index, like changing from tokenized to indexed or from Edm.String to Collection(Edm.String), ... or even re-index all the data with some additional information.
It would be great if we could have a "production" and a "staging" slot in an index (similar to Cloud Services for example) where we can do whatever we need to do in the staging slot, and as soon as we're ready (all data has been indexed), we simply swap the staging slot to the production slot and from that point on the new index is used.
…
418 votes -
Support hierarchical facets
Searching in a hierarchy is very common.
E.g. Cars/Ford/Accessories358 votesIn the meantime, please see this page for details on how to do this: http://blogs.technet.com/b/onsearch/archive/2015/09/09/multi-level-taxonomy-facets-in-azure-search.aspx
-
provide multiselect facets
When providing a multiselect on the frontend you will need a facet which also provides the non-selected values and their count. The count should than be based on all the other filters, but with the filter on it's own facet ignored.
For instance: You have a country facet which has 6 values:
USA, Canada, Mexico, Brasil, Chile, Argentina
When the user selects USA you send a filter: $filter=country eq USA.
But in the response your facet is now limited to just USA. But since you want to provide a multiselect you also need the other 5 countries in the facet…
343 votes -
Linq2AzureSearch
Beeing able to utilize LINQ in order to search in indexes just like LINQ2SQL looks like. The codebase would be simple to write, simple to read and analogous to LINQ2SQL query. Linq2Lucene project (https://github.com/themotleyfool/Lucene.Net.Linq) defines a very nice interface for fast time to market.
The Database context would be replaced with a Azure Search context that would wrap all REST based and JSON making the developer only focusing on the most important, Search.
Similar, delete and updates would be handled the same way a Linq2SQL handles it in the codebase.
Time2Market for Azure Searchbased inventions would be furious.
336 votes -
Support for enhanced security - including security trimming and AD integration
I suspect that having a search service with two levels of access (admin and user) isn't likely to be a great solution, particularly if it applies to all indexes within the service. For example, to be able to upload a document, you need the same key as someone that can create indexes. I was left with the feeling that there probably is a need for some sort of role-based security where users/keys can be associated with particular indexes within the service. Even better would be something that allowed for vertical and/or horizontal filtering at the level of each index. Is…
324 votes -
Case-insensitive sorting for string fields
When querying the index, is it possible to do a case-insensitive $orderby on a string field? For example, we want to sort results by username but preserve the user's choice of casing. While we could workaround this by keeping a normalized version of the username alongside the original data, having this feature built-in would obviously be preferable.
323 votes -
Search multiple indexes at once
The underlying elasticsearch technology supports making a single search query that searches and ranks results over multiple disparate indexes at once[1], but this functionality is not surfaced through the Azure Search APIs.
Adding this would allow a single search to span over indexes with different schemas combined in a single correctly ordered result set.
[1]: https://www.elastic.co/guide/en/elasticsearch/reference/current/multi-index.html
270 votes -
Possibility to delete all documents in an index without deleting and recreating the index
You should provide a possibility to delete all documents in an index without having to delete it and creating it again.
252 votes -
Support for Dictionary DataType
Dictionary datatype is very common. Often, we'd like to store our data as a dictionary. Currently, I serialize and store it as a string. It'd be great if the dictionary type is supported.
247 votes -
Pricing based on transactional throughput, similar to SQL Azure
The current model is a lot like the old SQL Azure models, where it was the volume of data that largely drove the cost of the service. SQL Azure became an affordable breakthrough when it shifted toward pricing based on overall transactional throughput (i.e., large databases that did not need to handle huge volumes were now affordable).
Search should work the same way. If I operate a marginally popular web forum, for example, I may have several gigs of data to index, but the storage of that is the cheap part. It doesn't get a lot of queries, and new…
218 votes -
Per index pricing model with no index cap
In my scenario I need to offer a single index per tenant. As my SaaS app grows I want to be able to add indexes as new tenants are on-boarded. Currently I have to purchase large blocks of indexes that will go unused until I on-board enough tenants. Then when I hit the cap I have to wither migrate to a new plan or (if I am past the cap of the largest plan) start partitioning tenants across multiple Search Services rather than scaling out from a single service to X number of private/isolated indexes.
207 votes -
Support for crawling HTML/websites
Enable Azure Search to crawl a local HTML website
198 votesAzure Search now supports crawling Azure blob storage and indexing HTML documents. While it’s not quite the same as crawling a website, this may be of significant help. Please see https://azure.microsoft.com/en-us/documentation/articles/search-howto-indexing-azure-blob-storage/ for more details.
Thanks,
Eugene Shvets
Azure Search -
'Skip' limit should be far higher - int.MaxValue
'Skip' is limited to 100,000 - which makes it impossible to iterate over an entire dataset, unless there's some meaningful way to segment the data set first.
Given there's an upper limit on the complexity of the queries we can pass in 4500 chars (it seems?)
It's very, very awkward to perform tasks on large complex datasets.
157 votes -
Allow customer to provide encryption key for indexed data
To meet enterprise security requirements, it would be great if index data could be encrypted (ideally with a key provided by the customer.)
138 votes
- Don't see your idea?