Custom Column Analyzers for Dataverse Search in Power Apps

Dataverse search uses Azure AI Search capabilities to return the best data based on what the user asks for. Custom column analyzers can be used to tailor the search results for specific table columns.

Overview

Dataverse search provides many different Azure AI Search capabilities, including index and search analyzers. These built-in capabilities help to define how an index maps certain phrases or words to the best match to data in a column and a row in a table.

When to Use Custom Column Analyzers

Custom column analyzers should be used when the default Azure AI Search analyzers do not provide the desired results. This can be due to the presence of special characters or frequently used words (such as 'of' and 'or') in the data.

Built-in Analyzers

By default, Dataverse Search uses the Microsoft language analyzer based on the Dataverse organization's base language. If there is no Microsoft analyzer, Dataverse search uses the Lucene Analyzer. Both indexed documents and search terms go through the analysis during indexing and query processing.

Availability of Built-in Search Analyzers

Azure provides a list of available search analyzers that can be used in custom column analyzers. These include language analyzers and keyword analyzers.

Creating Custom Column Analyzers

To create a custom column analyzer, a new row is created in the SearchAttributeSettings table in Power Apps. This requires a valid JSON file containing the custom analyzer definition.

Defining Custom Analyzers

Custom analyzers use the Azure AI Search library and are defined in a JSON file. The definition includes the tokenizer, token filters, and character filters used in the custom analyzer.

Uploading Custom Analyzers

Custom analyzers are uploaded to Power Apps, which then creates a new row in the SearchAttributeSettings table. Once the custom analyzer is uploaded, it can be assigned to a specific column.

Applying Custom Column Analyzers

Custom column analyzers are applied to a specific table column by creating a new row in the SearchAttributeSettings table. This row specifies the column and the custom analyzer to be used.

Adding Custom Column Analyzers to Dataverse Search

To add a custom column analyzer to Dataverse search, the custom analyzer needs to be uploaded to Power Apps. This is done by creating a new row in the SearchAttributeSettings table.

Steps to Add Custom Column Analyzers

1. Log in to Power Apps and navigate to the SearchAttributeSettings table.
2. Create a new row in the table and select the "Name" and "Attributename" columns to be visible.
3. In the "Settings" column, enter the JSON definition of the custom analyzer.
4. Save the row.
5. The custom analyzer is now enabled in Dataverse search.

Using Custom Column Analyzers with Dataverse Search

Once a custom column analyzer is created and uploaded, it can be used with Dataverse search. The custom analyzer is applied to a specific table column by creating a new row in the SearchAttributeSettings table.

Managing Custom Column Analyzers

Custom column analyzers can be managed by editing the rows in the SearchAttributeSettings table. This allows for the adding, modifying, or removing of custom column analyzers.

Examples of Custom Column Analyzers

Some examples of custom column analyzers include:

• Removing parentheses from the value of a column during indexing and search.
• Converting special characters to a standard format.
• Ignoring frequently used words (such as 'of' and 'or') in the data.

Conclusion

Custom column analyzers can be used to tailor the search results for specific table columns in Dataverse search. By creating and applying custom column analyzers, users can improve the accuracy and relevance of search results.