Data Wrangler
Cross-Functionality

General
Workflow

Canvas

When selecting exactly one table, two entries for Data Wrangler:

  1. Toolbar
  2. Preview Panel

Data Wrangler icon shows up for table that has operation history

When more than one tables are selected,
Data Wrangler should only be available from Preview Panel

Edit pipeline mode

Same rules on Canvas should also be applied in Edit Pipeline Mode

Scenario should consider the case when exiting the pipeline mode. Table icon and history should preserve

General workflow

  1. Load table
  2. Data Wrangler
  3. Cube publish

Union/Join

None-schema-change
Data Wrangler operation

Should not affect Union

Should only affect Join if the Join column is changed

Schema-changing
Data Wrangler operation

Validation rule for Union if user try to change schema. Check change data type scenarios

Validation rule for Join if user tries to change schema

All valid operation should be able to push down to the final table in the pipeline

Modeling

Incompatible datatype change should break linking

Rename column should make the old Object as unused and create a new Object

New derived column should have new Object mapping to it

Dropped column should unlink the Object

Data Wrangler change should not change Attribute/Metric type

Check behavior of interaction between multi-form and Data Wrangler

Facet

Check new derived column shows up the facet correctly

Check change data type changes the binning logic correctly

Check update the column changes the binning logic correctly

Edit&re-publish cube

Verify the operation history persists and the data grid is correct