Leveraging advanced knowledge to construct superior search functions with Azure Search

Information is never easy. Not each piece of information now we have can match properly right into a single Excel worksheet of rows and columns. Information has many various relationships such because the a number of places and telephone numbers for a single buyer or a number of authors and genres of a single e-book. After all, relationships sometimes are much more advanced than this, and as we begin to leverage AI to grasp our knowledge the extra learnings we get solely add to the complexity of relationships. For that cause, anticipating prospects to need to flatten the info so it may be searched and explored is commonly unrealistic. We heard this typically and it shortly turned our primary most requested Azure Search characteristic. Due to this we have been excited to announce the final availability of advanced sorts help in Azure Search. On this publish, I wish to take a while to clarify what advanced sorts provides to Azure Search and the sorts of issues you possibly can construct utilizing this functionality. 

Azure Search is a platform as a service that helps builders create their very own cloud search options.

What’s advanced knowledge?

Advanced knowledge consists of information that features hierarchical or nested substructures that don’t break down neatly right into a tabular rowset. For instance a e-book with a number of authors, the place every creator can have a number of attributes, can’t be represented as a single row of information except there’s a option to mannequin the authors as a group of objects. Advanced sorts present this functionality, they usually can be utilized when the info can’t be modeled in easy discipline constructions resembling strings or integers.

Advanced sorts applicability

At Microsoft Construct 2019,  we demonstrated how advanced sorts may very well be leveraged to construct out an efficient search utility. Within the session we seemed on the Journey Stack Alternate web site, one of many many on-line communities supported by StackExchange.

The StackExchange knowledge was modeled in a JSON construction to permit simple ingestion it into Azure Search. If we have a look at the primary publish made to this web site and deal with the primary few fields, we see that every one of them could be modeled utilizing easy datatypes, together with tags which could be modeled as a assortment, or array of strings.

   "id": "1",
    "CreationDate": "2011-06-21T20:19:34.73",
    "Rating": 8,
    "ViewCount": 462,
    "BodyHTML": "<p>My fiancée and I are on the lookout for a very good Caribbean cruise in October and have been questioning which
    "Physique": "my fiancée and that i are on the lookout for a very good caribbean cruise in october and have been questioning which islands
    "OwnerUserId": 9,
    "LastEditorUserId": 101,
    "LastEditDate": "2011-12-28T21:36:43.91",
    "LastActivityDate": "2012-05-24T14:52:14.76",
    "Title": "What are some Caribbean cruises for October?",
    "Tags": [
    "AnswerCount": 4,
    "CommentCount": 4,
    "CloseDate": "0001-01-01T00:00:00",​

Nonetheless, as we glance additional down this dataset we see that the info shortly will get extra advanced and can’t be mapped right into a flat construction. For instance, there could be quite a few feedback and solutions related to a single doc.  Even votes is outlined right here as a posh sort (though technically it might have been flattened, however that might add work to remodel the info).

"CloseDate": "0001-01-01T00:00:00",
    "Feedback": [
            "Score": 0,
            "Text": "To help with the cruise line question: Where are you located? My wife and I live in New Orlea
            "CreationDate": "2011-06-21T20:25:14.257",
           "UserId": 12
            "Score": 0,
            "Text": "Toronto, Ontario. We can fly out of anywhere though.",
            "CreationDate": "2011-06-21T20:27:35.3",
            "UserId": 9
            "Score": 3,
            "Text": ""Best" for what?  Please read [this page](
            "UserId": 20
            "Rating": 2,
            "Textual content": "What would you like out of a cruise? To loosen up on a ship? To go to islands? Tradition? Journey?
            "CreationDate": "2011-06-24T05:07:16.643",
            "UserId": 65
    "Votes": {
        "UpVotes": 10,
        "DownVotes": 2
    "Solutions": [
            "IsAcceptedAnswer": "True",
            "Physique": "That is lower than a solution, however greater than a remark…nnA giant share of your journey b
            "Rating": 7,
            "CreationDate": "2011-06-24T05:12:01.133",
            "OwnerUserId": 74

All of this knowledge is vital to the search expertise. For instance, you may wish to:

Actually, we might even enhance on the present StackExchange search interface by leveraging Cognitive Search to extract key phrases from the solutions to provide potential phrases for autocomplete because the person sorts within the search field.

All of that is now doable as a result of not solely are you able to map this knowledge to a posh construction, however the search queries can help this enhanced construction to assist construct out a greater search expertise.

Subsequent Steps

If you want to be taught extra about Azure Search advanced sorts, please go to the documentation, or try the video and related code I made which digs into this Journey StackExchange knowledge in additional element.

Show More

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button