Feed json array to Cognitive Search

Question

Hi all,

I do have a very simple json file which I have stored in a blob storage account.
My goal is to index it and eventually search through it.

I think I am missing some basic things here but I just cant figure it out ;)

Here's the json:

{
    "pages": [
        {
            "id": 1,
            "title": "Home",
            "url": "125715.htm",
            "text": "bla bla bla bla"
        },
        {
            "id": 2,
            "title": "Legal Notice",
            "url": "33205.htm",
            "text": "yada yada yada"
        }
    ]
}

The data source is set up and I can see the file.
An indexer is set up and I have changed the key to "id" instead of "metadata_storage_path" as the file is always the same (and therefore not unique) but I am trying to get each json array...

However, when saving, it sits at "validating..." forever. When I just save it with "metadata_storage_path" as key, the indexer is created instantaneously.

Any idea how I can achieve this?

Thanks!

Answer 1

To benefit the community/summarizing the comment discussion.

Thomas-4898 was able to resolve this issue by creating index, indexer and data source manually (but, not with the Import Data wizard).

Detailed steps below:

1.Indexing mode JSON Array
2.documentRoot: /pages

We currently do not have the option to set the document root in the portal. I have shared this feedback with our product team and is being looked into. However, there is no ETA to share at this time.

Check out the Parsing nested JSON Arrays section here.

@Thomas , thanks for the update and your patience!

-----------------------------------------------------------------------------------------------------------

To benefit the community find the right answers, please do mark the post which was helpful by clicking on ‘Accept Answer’ & ‘Up-Vote’.