Best practices to guarantee that the defined types in Power Query are carried forward into the Power BI Data Model are as follows:
Set Data Types Explicitly in Power Query: Always set the data type for each column in Power Query, even when it looks like it has been assigned automatically. By "transforming" into the Data Type dropdown, that data type can be specified. Specifying the data type tells Power Query that this is the anticipated treatment of the column in question through the transform process.
Immediate Data Type Setting: Immediately after importing or transforming the data that can change the content of the column, set the data types. Minimize possible changes due to intermediary steps within your query.
Check Column Types in the Data Model: Look at column properties in the Fields pane of the Data Model after data has been loaded in Power BI. Sometimes, Power BI will also assign types to data on loading in the case of abnormal characteristics in the columns of data being loaded. Correct any mismatched data types in the Data Model to ones that were expected.
Turn Off Auto Detection of Relationships: Although auto-detection by Power BI can create some of the most unexpected types, which will have been created as against the types expected from data's relationships, turn auto-detection relationships off through the Options menu and build and validate all relationships manually.
Continuous Date Coding: If in date columns, create a Date table, format it into the right configuration, and mark it out as a Date table in the model. This prevents the trouble in time intelligent functions for date evaluation as well as standardized treatment of date data.
