0.5.2 (2021-03-04)

Added

• Project template to be used as a starting point for projects intending to go into production.

• session.Session.start_transaction() to group a set of store operations in a single transaction.

• mode parameter to cube.Cube.query(). The "raw" mode uses a faster data transfer technique than the one used by the default "pretty" mode, leading to drastic performance improvements for larger query results (#146).

• include_totals parameter to cube.Cube.query() and query.cube.QueryCube.query() to indicate whether the returned DataFrame should include the grand total and subtotals.

• keep_totals parameter to session.Session.query_mdx() and query.session.QuerySession.query_mdx() to indicate whether the returned DataFrame should contain, if they are present in the query result, the grand total and subtotals.

• config.create_kerberos_authentication() to support Kerberos authentication providers.

User interface

• Convert to Widget Below JupyterLab command available on (unmutated) DataFrames returned by the cube.Cube.query() and query.cube.QueryCube.query() methods (#49).

  

  This command provides a bridge from queries run from the Python API to widgets offering more interactivity. It can be executed by:

  • right-clicking on the DataFrame and selecting the corresponding context menu item.

  • selecting it in the command palette when the DataFrame cell is the active one.

• Defer updates checkbox in the JupyterLab extension sidebar (#25).

  

  Deferring updates reduces the load on the server by avoiding intermediate queries.

• Pivot table columns can be resized (#206).

  • 

    Dragging the left border of a column resize the column to the desired width.

  • 

    Double clicking on the left border resize the column to fit its content.

  • 

    Selecting all the cells and double clicking on the left border of any column.

• Pivot tables support expansion by another hierarchy.

  

  Drag and drop a hierarchy to a pivot table cell to expand it.

Experimental

• atoti.experimental.create_date_hierarchy() to create a multilevel date hierarchy based on a date column.

Changed

• Improve where() performance when either:

  • no false_measure is passed.

  • the passed condition on levels is made only of == or != comparisons to None and at most one == or != comparison to a non-None value.

Fixed

• Issue when accessing query session hierarchies and levels (#202).

• Issue where switching to an incompatible widget type would crash the JupyterLab extension (#211).

• Issue with glob path on Windows.

• Prevent redefining special roles such as ROLE_USER and ROLE_ADMIN by raising an error in config.create_role().

• Respect sampling mode when adding rows with atoti.store.Store.append() and the += operator.