semantic-model-consumption

Imported from pugliabi/promptingbi (.cursor/skills/semantic-model-consumption/SKILL.md). Install upstream with npx skills add pugliabi/promptingbi --skill semantic-model-consumption. Copyright stays with the author.

Update Check — ONCE PER SESSION (mandatory) The first time this skill is used in a session, run the check-updates skill before proceeding.

• GitHub Copilot CLI / VS Code: invoke the check-updates skill.
• Claude Code / Cowork / Cursor / Windsurf / Codex: compare local vs remote package.json version.
• Skip if the check was already performed earlier in this session.

CRITICAL NOTES

1. To find the workspace details (including its ID) from workspace name: list all workspaces and, then, use JMESPath filtering
2. To find the item details (including its ID) from workspace ID, item type, and item name: list all items of that type in that workspace and, then, use JMESPath filtering

Power BI Semantic Model Consumption

Table of Contents

Prerequisites

• COMMON-CORE.md — Fabric concepts, authentication, and control-plane API context.
• COMMON-CLI.md — CLI-oriented discovery and token/audience patterns.

Must/Prefer/Avoid

MUST DO

• Call MCP tools for discovery and execution — use DiscoverArtifacts to find the semantic model GUID and ExecuteQuery to run DAX queries. These are MCP tool invocations, not shell commands. Do NOT use az rest, Get-Command, or PowerShell scripts to discover or invoke them.
• Keep this skill read-only: metadata discovery and analytical DAX queries only.
• Treat DAX data queries and INFO.VIEW.* as available to any user with read access to the semantic model; assume other INFO.* functions may require elevated permissions.
• Resolve workspace and semantic model item identity dynamically; do not hardcode IDs.
• Use DAX INFO.VIEW.* / INFO.* for metadata discovery before writing data queries.

PREFER

• Validate semantic model scope early (artifactId) before iterative query refinement.
• Discover semantic model schema progressively: use filtered and projected INFO.VIEW.* / INFO.* calls (e.g., SELECTCOLUMNS + FILTER) to fetch only the information directly relevant to the current task instead of retrieving the full schema up front. See discovery-queries.md § Narrowing Results (Projection + Filtering).
• Keep guidance provider-agnostic so tool endpoint migration is low risk.

AVOID

• Model-change operations in this skill.

Recommended Discovery Order

1. Run the discovery-queries.md § Scope Estimation Queries to estimate metadata scope (table, column, measure, and relationship counts) before deep discovery.
2. Start with INFO.VIEW.TABLES() for a fast table inventory.
3. Expand to INFO.VIEW.COLUMNS() and INFO.VIEW.MEASURES() for semantic details.
4. Use INFO.VIEW.RELATIONSHIPS() to validate joins and filter behavior.
5. Use the full query catalog in discovery-queries.md for deeper patterns.

Frequently Used INFO Functions

• INFO.VIEW.TABLES
• INFO.VIEW.MEASURES
• INFO.VIEW.COLUMNS
• INFO.VIEW.RELATIONSHIPS
• INFO.PARTITIONS
• INFO.MODEL
• INFO.STORAGETABLECOLUMNSEGMENTS
• INFO.DEPENDENCIES
• INFO.EXPRESSIONS
• INFO.ROLES
• INFO.STORAGETABLECOLUMNS
• INFO.CALCULATIONGROUPS
• INFO.CALCULATIONITEMS
• INFO.CULTURES
• INFO.OBJECTTRANSLATIONS
• INFO.USERDEFINEDFUNCTIONS
• INFO.REFRESHPOLICIES
• INFO.ATTRIBUTEHIERARCHYSTORAGES
• INFO.COLUMNPARTITIONSTORAGES
• INFO.COLUMNSTORAGES
• INFO.DICTIONARYSTORAGES
• INFO.HIERARCHYSTORAGES
• INFO.PARTITIONSTORAGES
• INFO.RELATIONSHIPINDEXSTORAGES
• INFO.RELATIONSHIPSTORAGES
• INFO.SEGMENTMAPSTORAGES
• INFO.SEGMENTSTORAGES
• INFO.STORAGEFOLDERS
• INFO.STORAGEFILES
• INFO.TABLESTORAGES
• INFO.GENERALSEGMENTMAPSEGMENTMETADATASTORAGES
• INFO.DELTATABLEMETADATASTORAGES
• INFO.PARQUETFILESTORAGES
• INFO.STORAGETABLES

Metadata Object → INFO Function Map

Query Execution

Use the FabricIQ MCP server tools directly — call them as MCP tool invocations:

How to find the artifact ID

1. Preferred: Call DiscoverArtifacts (MCP tool) with searchQuery set to the model name and artifactTypes: ["SemanticModel"].
2. Fallback only if MCP discovery is unavailable: Resolve via Fabric REST (az rest) — list items of type SemanticModel in the workspace and filter by displayName.

Executing a query

Call the ExecuteQuery MCP tool directly:

ExecuteQuery(
  artifactId = "<SEMANTIC_MODEL_GUID>",
  daxQueries = ["EVALUATE SUMMARIZECOLUMNS(...)"]
)

Getting the schema first

Call GetSemanticModelSchema(artifactId = "<SEMANTIC_MODEL_GUID>") to retrieve table/column/measure definitions before writing DAX.

Troubleshooting

• ExecuteQuery capability is unavailable in the MCP server
  • Issue: Query execution cannot start because ExecuteQuery is not available in the active tool list.
  • Cause: The FabricIQ MCP server is not registered, not loaded, or the current client session has stale tool metadata.
  • Fix: Verify the active MCP server/tool inventory and confirm ExecuteQuery is exposed.
• Advanced INFO functions return permission errors
  • Issue: Queries against INFO.* fail with authorization or privilege-related errors.
  • Cause: Many INFO.* functions require elevated semantic model permissions beyond standard read access.
  • Fix: Start with INFO.VIEW.* functions for read-oriented discovery.
• Metadata output volume is too large for focused analysis
  • Issue: Returning full metadata rowsets introduces too many properties and crowds the working context.
  • Cause: Unbounded INFO.VIEW.* and INFO.* queries return broad object/property surfaces that are often unnecessary for the current task.
  • Fix: Use the scope estimation queries in discovery-queries.md § Scope Estimation Queries to estimate scope and inspect output schemas, then narrow results with projection and filtering as shown in discovery-queries.md § Narrowing Results (Projection + Filtering).
• Do not use INFO DAX functions to retrieve role memberships
  • Issue: INFO.ROLEMEMBERSHIPS() returns empty or incomplete results.
  • Cause: Role members are assigned at the service level (Entra ID) after deployment, not in the model definition — so DAX INFO functions cannot reliably surface them.
  • Fix: Redirect to Manage role membership documentation for current guidance on role membership management.

Examples

For the full query catalog (including dependency patterns), see discovery-queries.md.

Sample Metadata Query

EVALUATE
INFO.VIEW.TABLES()
ORDER BY [Name]

Sample Data Query

DEFINE
MEASURE 'Sales'[Total Sales] = SUM('Sales'[Amount])
EVALUATE
SUMMARIZECOLUMNS(
    'Customer'[Customer Name],
    "Total Sales", [Total Sales]
)
ORDER BY [Total Sales] DESC