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
-
Support expressions in $orderby
Let us have an Azure Search index with three columns:
1. Id [Edm.String]
2. Tags [Collection(Edm.String)]
3. MaxScore [Edm.Int32]The index contains the following two entries:
Id Tags MaxScore
1 Paris,London,Rome 30
2 Paris 10We have also a scoring profile and a scoring function of type Tag (Sum, Linear) which evaluate every tag by 10 points (boost = 10).
We want to search the records with tags "Paris, London" in this index and sort it according to the percentage of compliance, while 100% is the value in column MaxScore.
The problem is that by default Azure Search sorts records…
50 votesThank you for your feedback. We’re considering this for a future release of Azure Search.
Thanks,
Liam
Azure Search Product Team -
Provide the ability to hibernate/revive the index
There are many scenarios where indexes are only needed for particular periods. A simple example is for demo purposes. I want to be able to have an index ready to use, but only use it periodically. It's too expensive to keep it running. It would be more 'cloudy' to have the ability to save/hibernate it to storage, and to be able to revive it from there. (I saw another suggestion to backup/snapshot. That's good but it also needs the ability to be deleted and still be recovered).
47 votes -
Offer reserved pricing on Azure Search (similar to AWS ElasticSearch)
AWS offers up to 52% discount when committing to a 3 year reserved instance for AWS ElasticSearch.
There is no reserved pricing option for Azure Search.
Please provide reserved pricing, similar to Cosmos DB/Azure SQL, at a discount comparable to AWS.
47 votes -
Prospective search (aka percolator or reverse search)
It would be huge for my application to be able to do reverse searches a la Elasticsearch's percolator feature: Given a document ID, which saved queries would match that document.
44 votes -
Random Scoring
See here https://www.elastic.co/guide/en/elasticsearch/guide/current/random-scoring.html
This has come up in the past on SO by other users as well.44 votesThank you for your feedback. While it is unlikely we’ll address this suggestion in the near future, we’ll reassess based on the number of votes it receives. The workaround mentioned here is very cool, by the way!
Thanks,
Mike
Azure Search Product Team -
Disable or replace TF-IDF for relevance computing
(Posting on behalf of Xu Lan)
To fit the needs of our search application, we would like to change the relevance computing by disabling or replacing TF-IDF.
Elasticsearch provides a mechanism to ignore TF-IDF: https://www.elastic.co/guide/en/elasticsearch/guide/current/ignoring-tfidf.html#ignoring-tfidf.44 votesThank you for your feedback. While it is unlikely we’ll address this suggestion in the near future, we’ll reassess based on the number of votes it receives.
Thanks,
Luis Cabrera
Azure Search Product Team -
Clone entire Azure Search instance, just like I can clone App Service.
I would be extremely handy if I could clone entire Azure Search instance to the same or a new Resource Group, in the same or a new location using the same or another Azure Subscription. Like I can clone Azure App Service wherever I want.
User case: If there is an issue which is reproducible with production instance of Azure Search, then I will clone in to a separate sandbox, and take it from there. As of today, I have to create a brand new Azure Search instance, clone SQL Database and reindex them to get the same data.
42 votesThank you for your feedback. While it is unlikely we’ll address this suggestion in the near future, we’ll reassess based on the number of votes it receives.
Thanks,
Luis Cabrera
Azure Search Product Team -
Support picking existing index when creating an indexer in Azure Portal
Today is only possible to create an indexer if you follow the entire process through import data workflow:
-Select the data Source
-Add Cognitive skills (optional)
-Create an Index
-Create an IndexerHowever, there are times we want to set a different schedule for an indexer or create a new one, without recreating the index.
41 votesThank you for your feedback. While it is unlikely we’ll address this suggestion in the near future, we’ll reassess based on the number of votes it receives.
Thanks,
Luis Cabrera
Azure Search Product Team -
Add support for inheritance ($type)
Sometimes we have inheritance in our database tables. In my case, I have inheritance mapped to a table using EF Code First, and the table contains a generated column, [Discriminator], that tells the EF which type should be activated when fetching the entity.
There should be a way for Azure Search (a field transformation function maybe) to map discriminators in the database to fully qualified non-searchable $type fields in the indexes, in order for the SDK to be able to deserialize the types correctly using the TypeNameHandling setting of the Json.net serializer.
40 votesThank you for your feedback. While it is unlikely we’ll address this suggestion in the near future, we’ll reassess based on the number of votes it receives.
Thanks,
Vinod
Azure Search Product Team -
Support filter min score
Elasticsearch support "min_score" to filter documents by scoring.
https://www.elastic.co/guide/en/elasticsearch/reference/current/search-request-min-score.htmlCan you add this feature to allow us to deal with irrelevant results please ?
40 votesThank you for your feedback. While it is unlikely we’ll address this suggestion in the near future, we’ll reassess based on the number of votes it receives.
Thanks,
Elad
Azure Search Product Team -
Please provide support for MySQL database for pushing data to the index from azure portal
Please provide support for MySQL database for pushing data to the index from azure portal
39 votesThis feature is currently in private preview (REST API support only, not portal support yet). For onboarding instructions, please email Victor Liu (victliu at the usual Microsoft domain)
-
Provide option to return total facet counts without all facet values included in response
We have facets with thousands of unique values, sometimes up to around 100k. In our UI, we'd like to show the top 10 or so values of each facet to allow the user to filter by them, but we have no need to display all 100k. That said, we need to display the facet counts so they know roughly how much data is being searched.
To do that currently, we have to set the facet counts arbitrarily high to make sure the accurate counts are retrieved, but that means we're getting back huge HTTP responses when really we just want…
39 votesThank you for your feedback. While it is unlikely we’ll address this suggestion in the near future, we’ll reassess based on the number of votes it receives.
Thanks,
Vinod
Azure Search Product Team -
Support scoring profiles on suggesters
I think it would be valuable to expand the capabilities of suggesters. Right now, they're somewhat limited. I think it would be nice if they provided the flexibility of scoring profiles. The reason why is because sometimes, it might be nice to show the most recently added or viewed documents (i.e. Edm.DateTimeOffset) as suggestions. Other times, you may want to show the most popular, or most viewed, documents (i.e Edm.Int32). These are just two scenarios that I would like to cover with suggestions, but, I'm a bit limited.
38 votesThank you for your feedback. This is an interesting suggestion, but it is unlikely we’ll address this in the near future. We’ll reassess based on the number of votes it receives.
Thanks,
Mike
Azure Search Product Team -
Increase Blob Indexer file size limits
The "Blob indexer: maximum characters of content extracted from a blob" limit seems rather restrictive, 32K for free, 64K for standard and 4MB for everything else. Plus when this limit is reached all you see in the azure portal is
[{ "key": null, "message": "Truncated extracted text to 32768 characters." }]
It'd be useful to know which documents are affected.
37 votesThank you for your feedback. We’re considering this for a future release of Azure Search.
Thanks,
Luis Cabrera
Azure Search Product Team -
Language analyzer - ability to disable word decompounding
Azure Search has word decompounding (in languages like German, Danish, Dutch, Swedish, Norwegian, Estonian, Finish, Hungarian, Slovak)
https://docs.microsoft.com/en-us/rest/api/searchservice/language-supportHowever often it doesn't lead to desired output as we need to find only compound words.
Especially interested to have it in "et.microsoft" analyzer.36 votesThank you for your feedback. While it is unlikely we’ll address this suggestion in the near future, we’ll reassess based on the number of votes it receives.
Thanks,
Luis Cabrera
Azure Search Product Team -
35 votes
We are actively evaluating this feature.
-
Support IP firewall\service endpoints
Our enterprise requires 2 factor auth for all remote services. Please support IP based firewall and service endpoints!
34 votesThank you for your feedback. We’re considering this for a future release of Azure Search. Will supporting V-NETs for Azure Search be a good solution for your use case?
Thanks,
Elad
Azure Search Product Team -
Change hit higlighting to handle line breaks as well as sentences
It looks like the highlighting feature returns the full sentence around the highlighted term. This makes sense and works when the content has full stops in it.
We are indexing documents which have bullet points and lists which do not have a full stop. As a result Azure search returns a large amount of content around the highlighted term.
33 votesThank you for your feedback. We’re considering this for a future release of Azure Search.
Thanks,
Mike
Azure Search Product Team -
sintax highlight length
Could we get a parameter to specify the length of a highlight fragment that is returned when the highlight parameter is used on a field? The string returned is too short for some of our users because they are searching very techincal data and need more context. It would be very useful to be able to return longer fragments depending on the field we're searching so we need control over this.
30 votes -
Indexer execution status to App insights
Scenario: Thanks to the power of Azure, I have setup a indexer which continuously updates my index from the Azure sql db. Now I want to know if there are any errors while processing any document and why it failed to update the index(error message). This information is very critical for new documents or any document which is modified in the prod database. Note: Different teams do not have access to the Azure search service portal.
Current state: We already have indexer execution status but this information needs to be processed by another process by polling.
Ask: If this information…
30 votesI wonder if this article meets your needs:
https://docs.microsoft.com/en-us/azure/search/search-monitor-usageYou can send operation logs to log analytics to allow you to undestrand the health of your service.
Thanks!
Luis Cabrera
Azure Search Product Team
- Don't see your idea?