TabularEditor/TabularEditor 2.21.0

Tabular Editor 2.21.0

on GitHub

• Windows installer: TabularEditor.2.21.0.Installer.msi
  • SHA256: 0E79FF597F401E8943583FCE633ABD196BD1A3BFA423B2DB537ED93010458A8E
• Portable version: TabularEditor.2.21.0.zip
  • SHA256: EE018E3FCB82D87F1D8787807713026D903A28A78792489FDF8D5C03EA185650

Improvements in 2.21.0

• Updated AMO/TOM to 19.69.6.2. This adds support for the latest TOM modelling features. For example:

  • For DirectLake models, we now support the DirectLakeBehavior property, which specifies whether the model should automatically fallback to DirectQuery mode (default), or be restricted to only DirectLake mode, or only DirectQuery mode.

  • Support for Calculation Group Selection Expressions (the model needs to be at Compatibility Level 1605 for these properties to be available).

• Updated TMDL to the latest preview version.
  Note: TMDL is a preview feature. Things are still subject to change.
• When the Timeout property of a Provider (legacy) Data Source has been specified (i.e. a nonzero value), Tabular Editor will now use the value for the CommandTimeout property when performing table imports/schema updates, see #1115 (thanks @rbulka)
• The C# .Tokenize() method (used to retrieve a list of DAX tokens for objects that contain DAX expressions, such as measures, calculation items, etc.), now take an optional bool includeHidden argument. By default, this is true. Set this flag to false if you do not want whitespace and comment tokens in the output. This method should be used in dynamic LINQ expressions, if you want to create BPA rules that check for the presence of certain DAX keywords or syntax. It is more robust than e.g. Regex. See examples here.

Bugfixes in 2.21.0

• When deploying a model, always add a dummy partition to an incremental refresh table, if the metadata does not define any partitions. Otherwise, the generated TMSL deployment script would be invalid.
• Fixed a bug that could cause a NullReferenceException when deleting objects through a C# script, see #1139
• The data source name is now (again) shown under the source info-column for tables that use Provider Data Sources, see #1136
• Fixed an issue that prevented loading a model from a Power BI workspace with special characters in the name, see #1133
• Fixed a bug where the Best Practice Analyzer would report the wrong number of objects having a violation within each rule.