Step 5: Querying the Amazon Kendra index

Your Amazon Kendra index is now ready for natural language queries. When you search your index, Amazon Kendra uses all the data and metadata you provided to return the most accurate answers to your search query.

There are three kinds of queries that Amazon Kendra can answer:

• Factoid queries ("who", "what", "when", or "where" questions)

• Descriptive queries ("how" questions)

• Keyword searches (questions whose intent and scope are not clear)

Querying your Amazon Kendra index

You can query your Amazon Kendra index using questions that correspond to the three kinds of queries that Amazon Kendra supports. For more information, see Queries.

The example questions in this section have been chosen based on the sample dataset.

To query your Amazon Kendra index (Console)

1. Open the Amazon Kendra console at https://console.aws.amazon.com/kendra/.

2. From the Indexes list, click on kendra-index.

3. From the left navigation menu, choose the option to search your index.

4. To run a sample factoid query, enter Who is Lewis Hamilton? in the search box and press enter.

   The first returned result is the Amazon Kendra suggested answer, together with the data file containing the answer. The rest of the results form the set of recommended documents.

   

5. To run a descriptive query, enter How does Formula One work? in the search box and press enter.

   You will see another result returned by the Amazon Kendra console, this time with the relevant phrase highlighted.

   

6. To run a keyword search, enter Formula One in the search box and press enter.

   You will see another result returned by the Amazon Kendra console, followed by the results for all other mentions of the phrase in the dataset.

   

To query your Amazon Kendra index (AWS CLI)

1. To run a sample factoid query, use the query command:

   Linux

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "Who is Lewis Hamilton?" \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   macOS

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "Who is Lewis Hamilton?" \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   Windows

   aws kendra query ^
           --index-id kendra-index-id ^
           --query-text "Who is Lewis Hamilton?" ^
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   The AWS CLI displays the results of your query.

2. To run a sample descriptive query, use the query command:

   Linux

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "How does Formula One work?" \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   macOS

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "How does Formula One work?" \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   Windows

   aws kendra query ^
           --index-id kendra-index-id ^
           --query-text "How does Formula One work?" ^
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   The AWS CLI displays the results to your query.

3. To run a sample keyword search, use the query command:

   Linux

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "Formula One" \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   macOS

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "Formula One" \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   Windows

   aws kendra query ^
           --index-id kendra-index-id ^
           --query-text "Formula One" ^
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   The AWS CLI displays the returned answers to your query.

Filtering your search results

You can filter and sort your search results using custom document attributes in the Amazon Kendra console. For more information on how Amazon Kendra processes queries, see Filtering queries.

To filter your search results (Console)

1. Open the Amazon Kendra console at https://console.aws.amazon.com/kendra/.

2. From the Indexes list, click on kendra-index.

3. From the left navigation menu, choose the option to search your index.

4. In the search box, enter Soccer matches as a query and press enter.

5. From the left navigation menu, choose Filter search results to see a list of facets you can use to filter your search.

6. Select the check box for "Champions League" under the EVENT subheading, to see your search results filtered only by the results containing "Champions League".

   

To filter your search results (AWS CLI)

1. To see the entities of a specific type (such as EVENT) that are available for a search, use the query command:

   Linux

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "Soccer matches" \
           --facets '[{"DocumentAttributeKey":"EVENT"}]' \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   macOS

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "Soccer matches" \
           --facets '[{"DocumentAttributeKey":"EVENT"}]' \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   Windows

   aws kendra query ^
           --index-id kendra-index-id ^
           --query-text "Soccer matches" ^
           --facets '[{"DocumentAttributeKey":"EVENT"}]' ^
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   The AWS CLI displays the search results. To get a list of facets of type EVENT, navigate to the "FacetResults" section of the AWS CLI output to see a list of filterable facets with their counts. For example, one of the facets is "Champions League".

   Note

   Instead of EVENT, you can choose any of the index fields you created in Creating an Amazon Kendra index for the DocumentAttributeKey value.

2. To run the same search but filter only by the results containing "Champions League", use the query command:

   Linux

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "Soccer matches" \
           --attribute-filter '{"ContainsAny":{"Key":"EVENT","Value":{"StringListValue":["Champions League"]}}}' \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   macOS

   aws kendra query \
           --index-id kendra-index-id \
           --query-text "Soccer matches" \
           --attribute-filter '{"ContainsAny":{"Key":"EVENT","Value":{"StringListValue":["Champions League"]}}}' \
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   Windows

   aws kendra query ^
           --index-id kendra-index-id ^
           --query-text "Soccer matches" ^
           --attribute-filter '{"ContainsAny":{"Key":"EVENT","Value":{"StringListValue":["Champions League"]}}}' ^
           --region aws-region

   Where:

   • kendra-index-id is your saved kendra-index-id,

   • aws-region is your AWS region.

   The AWS CLI displays the filtered search results.