Task 2.2: Copy data

• ASwitch the destination to automatic settings so that the mapping adjusts itself

  Incorrect

  Automatic settings aren't available for Warehouse destinations, and they wouldn't make an unsupported type loadable.

• BChange the schema option from fixed schema to dynamic schema

  Incorrect

  With a warehouse destination only fixed schema is supported, and the schema option doesn't change which types load.

• CTurn on staging for the query so the warehouse compute converts the type

  Incorrect

  Staging changes where transformations run; the destination still can't store a DateTimeZone column.

• DConvert OrderedAt to Date/Time in the query before the destination

  Correct

  The Fabric Warehouse destination doesn't accept the DateTimeZone type, so the column must reach it as a Date/Time value.

Dataflow Gen2 destinations support different data types. For a Fabric Warehouse destination, DateTimeZone, Duration, Currency, Binary and some others aren't supported, so such columns must be converted in the query, for example DateTimeZone to DateTime. Automatic settings are supported for Lakehouse, Azure SQL Database and Snowflake destinations, and when loading into a warehouse only fixed schema is supported.

• AHave Nadia update the notebook's schedule

  Correct

  A scheduled notebook runs under the identity of whoever created or last updated the schedule, so Nadia's update moves runs to her account.

• BHave Nadia run the notebook once on demand

  Incorrect

  A manual run uses her identity for that run only; scheduled runs keep the identity of whoever last updated the schedule.

• CGive Nadia the Admin role in the workspace

  Incorrect

  Her permissions are already enough; runs keep using Omar's identity until the schedule is updated.

• DAttach the notebook to a different environment

  Incorrect

  The environment controls libraries and Spark settings, not the identity of scheduled runs.

A notebook run that the scheduler triggers runs under the identity of the user who created or last updated the schedule. Before creating or updating a schedule, make sure that user has the permissions every operation in the notebook needs. When that identity is removed, a user with the right access must update the schedule so that later runs use their identity.

• AAdd a filter step so the query folds to the source

  Incorrect

  The query already has no transformations; the engine choice comes from the row-count threshold, not folding.

• BTurn off staging for the query

  Incorrect

  Staging doesn't decide whether fast copy is used for a source this small.

• CTurn on Require fast copy for the query

  Correct

  Require fast copy bypasses the size threshold, and the refresh stops if fast copy can't be used.

• DChange the destination to a Fabric warehouse

  Incorrect

  Fast copy currently loads directly only to a lakehouse; a warehouse destination would need staging and a second query.

With fast copy allowed, a dataflow switches to the pipeline Copy activity backend automatically when the data passes a threshold: 100 MB of CSV or Parquet files in ADLS Gen2 or Blob storage, or 5 million rows in a database source. You can bypass the threshold by selecting Require fast copy, which also stops the refresh if fast copy can't be used. Fast copy currently loads directly only to a Lakehouse destination.

• AOnly extract new data when the maximum value in this column changes

  Incorrect

  That required setting decides when a bucket is refreshed; it still loads the current, incomplete month.

• BRequire incremental refresh query to fully fold

  Incorrect

  It validates that the query folds to the source; it doesn't hold back the current month.

• COnly extract data for concluded periods

  Correct

  With a monthly bucket size, this setting extracts only complete months that contain no future data.

• DExtract data from the past: 1 month

  Incorrect

  That setting controls how far back the initial load reaches, not whether an incomplete month is loaded.

Dataflow Gen2 incremental refresh divides data into buckets on a Date, DateTime or DateTimeZone column and refreshes a bucket when the maximum value of the change-detection column changes. The optional setting Only extract data for concluded periods limits extraction to periods that are complete, for example whole months with a monthly bucket size. Extract data from the past sets the range of the initial load.

• AINSERT INTO dbo.Orders SELECT * FROM RetailLH.dbo.orders

  Incorrect

  Three-part names reach tables of items in the same workspace, not CSV files in the Files section of a lakehouse in another workspace.

• BCOPY INTO dbo.Orders FROM 'https://onelake.dfs.fabric.microsoft.com/<workspaceId>/<itemId>/Files/orders/*.csv'

  Correct

  COPY INTO can load files from OneLake directly, with a path that uses the workspace and item IDs and the user's Microsoft Entra identity.

• CCOPY INTO dbo.Orders FROM 'https://onelake.dfs.fabric.microsoft.com/Retail/RetailLH.Lakehouse/Files/orders/*.csv'

  Incorrect

  With OneLake as the source, paths must use workspace and item IDs; friendly workspace and item names aren't supported.

• DCOPY INTO dbo.Orders from the ID-based OneLake path, with a shared access signature in the CREDENTIAL clause

  Incorrect

  For OneLake sources, only Microsoft Entra ID and the workspace identity are supported; SAS tokens, shared keys and connection strings aren't.

The COPY statement in a Fabric warehouse loads CSV, JSONL and Parquet files from Azure storage accounts and from OneLake, so files in a lakehouse's Files section can be ingested without an external staging account. With OneLake as the source, the path must use workspace and item IDs, the credential must be Microsoft Entra ID or the workspace identity, and a user who runs it with their own identity needs at least the Contributor role on both the source and the target workspace. T-SQL that reads tables with three-part names works only within the same workspace.

• APublishing moves the next scheduled refresh to the publish time

  Incorrect

  The schedule stays as it is; the extra refresh is the on-demand one that publishing starts.

• BThe editor keeps refreshing the data preview after the dataflow is published

  Incorrect

  Data previews run while you author; the queries seen after publishing come from the refresh it starts.

• CThe 02:00 refresh failed, and failed refreshes retry at the next publish

  Incorrect

  Nothing ties retries to publishing; a successful publish itself starts the refresh.

• DA successful publish starts an on-demand refresh of the dataflow

  Correct

  When a dataflow publish completes successfully, an on-demand refresh starts.

A dataflow refresh can be triggered on demand or by a refresh schedule of up to 48 slots a day. Besides selecting Refresh or running a pipeline with a Dataflow activity, a successful publish of a dataflow also starts an on-demand refresh. Plan publishes of dataflows that read busy sources with that in mind.

• ASet the end date far in the future, such as 2099

  Correct

  Pipeline schedules need both a start and an end date, so a distant end date keeps it running long term.

• BLeave the end date empty to create an open-ended schedule

  Incorrect

  There's no option for an open-ended schedule.

• CSet the end date to today and turn on auto-renew

  Incorrect

  Schedules have no auto-renew; they end on their end date.

• DUse an event-based trigger instead, which has no end date

  Incorrect

  An event trigger starts runs on events, not daily at a set time.

When you schedule a pipeline, you specify a schedule type, frequency, start and end dates and times, and a time zone. You must set both a start and an end date because there's no option for an open-ended schedule; to keep a pipeline running long term, set the end date far in the future, for example 01/01/2099. You can edit or stop the schedule at any time.

• AQuery

  Incorrect

  Query is for statements that return a result set, such as SELECT.

• BNonQuery

  Correct

  NonQuery runs statements that don't return data, such as INSERT, UPDATE, DELETE, CREATE and DROP.

• CStored procedure

  Incorrect

  That's a separate activity, not a Script activity script type.

• DLookup

  Incorrect

  Lookup is another activity that reads data; it isn't a Script activity setting.

The Script activity runs query and nonquery SQL scripts against data stores that support DML and DDL, such as Fabric Data Warehouse. In its settings you choose Query to get a data result from SELECT statements or NonQuery for statements that don't return data, such as INSERT, UPDATE, DELETE, CREATE and DROP.

• AOwnership of every connection the dataflow uses

  Incorrect

  Access to the connections is required, not ownership, and she already has access.

• BA Power BI Pro license

  Incorrect

  The documented requirement is the workspace role and connection access, not a Pro license.

• CEdit permission shared on the dataflow item

  Incorrect

  Item-level edit permission doesn't replace the Member role that refreshing requires.

• DThe Member role or higher in the workspace

  Correct

  The user who triggers a dataflow refresh must have Member or higher access to the workspace and access to its connections.

To refresh a Dataflow Gen2, the user who triggers the refresh must have Member (or higher) access to the workspace and access to all connections the dataflow uses. The Contributor role doesn't meet that prerequisite, and neither a license nor an item-level permission replaces it.

• AAdd each one with + New, typing its name, type and nullability

  Incorrect

  That works, but it's the manual route; Import is quicker.

• BDefine them as pipeline variables with the same names

  Incorrect

  Pipeline variables don't define the activity's procedure parameters.

• CSelect Import to pull them from the procedure

  Correct

  Import brings in the parameters with their names and types, ready for values.

• DWrap the call in a Script activity instead

  Incorrect

  A Script activity can run the call, but it doesn't import the procedure's parameter definitions.

In the Stored procedure activity's settings, after choosing a connection and a stored procedure, you can select Import to import the parameters from the stored procedure as it's stored in the data source, or add them manually with + New by providing name, data type and nullability. Each parameter then takes a value or a dynamic expression.