Improved the performance of connecting to text files in shared network folders.
When you connected to PostgreSQL, and used a modulo operator in a DATEPART calculation, an error would occur: "ERROR: operator does not exist: double precision integer; Error while executing the query"
If your Preferences.tps file contained unexpected characters, adding a field to Color in the Marks card would cause Tableau Desktop to crash or throw a "No such function" error.
In certain environments, 64-bit Tableau Desktop would crash at the splash screen and fail to launch.
In some cases, when connecting to SAP HANA, certain variables and input parameters would be duplicated and attempting to define values for them would result in an internal error.
When you performed a union of Google Sheets documents, the merge mismatched fields functionality was missing.
Attempting to edit certain table calculations would result in an error: "Invalid input: expected '['"
Publishing workbooks containing multiple data sources and string fields converted to date would sometimes fail with a "No such function DATETRUNC" error.
When you opened and closed a second instance of Tableau Desktop, your first instance would be signed out of Tableau Server if it was previously connected.
The Tableau Online Sync Client would sometimes error or crash when signing in.
It was not possible to connect to a PostgreSQL database using a password containing certain combinations of curly brace and plus sign characters.
Mitigated an issue with creating extracts on the unsupported Windows Vista and Windows Server 2008 (non-R2) operating systems. Beginning with Tableau 10.3, we will begin blocking installations on these operating systems.
In some cases, creating an aggregated extract would fail with an "EncodeInteger: bad argument" error.
Editing a local copy of a published Google Analytics data source in Tableau Desktop would sometimes result in an error that prevented the data source from being edited: "Tableau/Missing required context variable: account-id"
When you created a histogram using a continuous bin, and fixed the bin axis using float values, the axis would revert to integer values when re-opening the workbook.
In some cases, colors assigned based on a group field would revert to default when you extracted the data source.
When creating a data source with multiple connections, replacing an SAP HANA table with an identically named table from another connection would result in an internal error.
When connected to a Microsoft SQL Server table with Thai_CI_AS collation, dimension members that only differed by space or period characters would be combined.