​
Direct Query

Power BI can connect directly to Honeydew in Direct Query mode by using the Honeydew ODBC driver. A Honeydew domain will be presented to the data model as a single virtual table.
Excel on Honeydew runs using Power BI connection. See Excel for details.

​
Power BI Service

Power BI Service can connect to Honeydew in Direct Query mode using the Power BI Gateway.

​
Metadata Sync

With Metadata Sync, Honeydew synchronizes the following metadata to Power BI:
  • Table display names (based on domain and entity display names). When no display name is set, Honeydew creates an automatic user friendly name.
  • Field type (metric or attribute) - attributes have default aggregation set to “none” and are presented as fields. regardless to data type, while metrics have a default aggregation set and are presented as measures.
  • Field display names. When no display name is set, Honeydew creates an automatic user friendly name.
  • Field folders. When no folder is set, Honeydew uses entity names as folder names.
You can create multi-level nested folders in Power BI by placing / separators in the folder value in Honeydew metadata (e.g. folder/subfolder)
  • Field hidden property.
  • Field descriptions.
  • Field format and semantic role (through metadata configuration).
Attributes have “none” as a default aggregation to distinguish them from metrics. But aggregating on attributes is possible! Any specific Power BI widget that needs to aggregate can do so by choosing an aggregation in the widget “field well”.

​
Power BI-specific Metadata

Attributes and Metrics: You can add Power BI-specific metadata, that specifies the default formatting or the semantic role, to an attribute or a metric in the schema: 
...
metadata:
  - name: powerbi
    metadata:
      - name: formatstring
        value: yyyy-mm-dd
      - name: datacategory
        value: Latitude
Some examples for formatstring values (Power BI DAX documentation):
  • yyyy-mm-dd
  • #,##0;-#,##0
Some examples for datacategory values:
  • Latitude
  • Longitude
  • PostalCode
  • City
  • StateOrProvince
  • Continent
  • Place
  • WebUrl
  • ImageUrl
  • Barcode
Domain: By default, Honeydew metadata sync creates a separate default folder for each entity within the domain. You can add Power BI-specific metadata to disable this behavior, so only attributes and metrics explicitly placed in folders appear in the folder structure. 
...
metadata:
  - name: powerbi
    metadata:
      - name: create_default_entity_folder
        value: false

​
Cross-Filtering and Referential Integrity

Honeydew supports cross filtering similar to Power BI. The following table helps translate Power BI settings to Honeydew:
Power BI Referential IntegrityPower BI Cross-FilteringHoneydew Join Type (from many to 1)Honeydew Cross-Filtering
No (Power BI default)Single (Power BI default)LEFTOne-to-many
NoBothLEFT (Honeydew default)Both (Honeydew default)
YesSingleINNEROne-to-many
YesBothINNERBoth
Note that that the default relationship settings in Honeydew are different than default Power BI settings.

​
Data Caching in Power BI

Honeydew supports caching the attributes of entities within Power BI in order to speed up queries over those entities. Caching is done by exposing cached entities as separate “virtual tables” in Power BI with a storage mode.
Power BI entity caching is helpful to keep in-memory small dimensions (up to 1GB in total size, up to 1M rows per each entity).For general purpose attribute caching in Snowflake see Entity Caching
Data caching in Power BI is supported for:
  1. Attributes only.
  2. Dimensions, not facts: Entities that are not the top granularity, and have a single join path to a higher granularity.
  3. With one-to-many or none cross filtering
Metrics are not cached in Power BI - use features such as Aggregate Awareness to speed up aggregations.
To set up Power BI data caching add the following to an entity YAML 
type: entity
# ... entity configuration
metadata:
  - name: powerbi
    metadata:
      - name: semantic_model_as_separate_table
        value: yes
      - name: semantic_model_storage_mode
        value: dual
Dual Storage Mode enables Power BI Service to cache entity data.Attribute queries using that entity will be answered in-memory. Other queries (such as metrics or combining the entity with other entities) will be translated in Honeydew with a direct mode.
Supported storage modes:
  • Not set: entity will appear as a folder under the main domain virtual table.
  • Set: entity will appear as a separate virtual table.
    • directQuery: No performance impact on queries. The entity appears as a separate virtual table.
    • dual: Speeds up queries that only need entity attributes. There is no performance impact on other queries. Use for dimensions with 1M rows or less.
    • import: Can negatively impact the performance of queries that join across entities. Not recommended - when in doubt, always prefer dual.
Refreshing data in Power BI caches is the user’s responsibility. Scheduled refresh can be set up either with Honeydew, or in Power BI Service.To refresh using Honeydew, use Metadata Sync APIs. Every metadata sync triggers a data refresh.

​
Limitations

DAX limitations on a Direct Query apply to a Honeydew connection. Logic that would otherwise require advanced DAX can be built directly in Honeydew instead. That includes complex multi-step calculations, filtered or partial metrics, calculated fields, time intelligence and other advanced features. Every Honeydew metric is consumable with Power BI in a Direct Query mode. Refer to relation modeling on how to model 1:1 or many:many relations which are not directly supported. A 1:1 Power BI relationship with a preset cross-filtering direction typically maps to a Honeydew 1:many relation with same direction.

​
Case Sensitive Data

Power BI is case-insensitive when handling data values. This means that if a column contains values with different capitalizations (e.g., name, Name, and NAME), Power BI may randomly choose one spelling and align all other values to that spelling. When used as a filter, it might only send the selected spelling, potentially omitting other matching data.To prevent this, the best practice is to standardize the column’s case (e.g., converting all values to uppercase) within the data definition in the Honeydew source.

​
Import

Power BI can import data from dynamic datasets in Snowflake.

​
With a VIEW or a TABLE

Dynamic datasets can be set up deployment to Snowflake. The resulting VIEW or TABLE can be used for a Data Import.

​
With Snowflake Native Application

Create a blank query, and open advanced query editor to write the query below: To read from a dynamic dataset: 
let
    Source = Value.NativeQuery(
            Snowflake.Databases(
                "SnowflakeHost.snowflakecomputing.com",
                "VirtualDataWarehouse",
                [Role="Role"]
            ){[Name="Database"]}[Data],
            "
            CALL SEMANTIC_LAYER_ENTERPRISE_EDITION.API.SELECT_FROM_DYNAMIC_DATASET(
                'workspace',
                'branch',
                'dataset');
            ",
            null,
            [EnableFolding=false]
    )
in
    Source
Or create an ad-hoc semantic query: 
let
    Source = Value.NativeQuery(
            Snowflake.Databases(
                "SnowflakeHost.snowflakecomputing.com",
                "VirtualDataWarehouse",
                [Role="Role"]
            ){[Name="Database"]}[Data],
            "
            CALL SEMANTIC_LAYER_ENTERPRISE_EDITION.API.SELECT_FROM_FIELDS(
                -- workspace / branch
                'workspace', 'branch',
                -- domain (can be NULL)
                'domain',
                -- dimensions - can be ad-hoc expressions
                ['entity.attr', ...],
                -- metrics - can be named or ad-hoc expressions
                ['entity.metric', 'SUM(entity.measure)', ...],
                -- filters - can be named or ad-hoc expressions
                ['entity.filter', 'entity.attr > 0', 'entity.name like ''%cheese%''', ...]
            );
            ",
            null,
            [EnableFolding=false]
    )
in
    Source
Given Snowflake connection parameters (host, warehouse, role and database) and Honeydew workspace, a working branch in it (usually would be prod), and a dynamic dataset dataset. This calls a stored procedure to retrieve dynamic dataset data to Power BI.