It was not possible to connect to some Google Sheets with special characters in the document name.
When you customized the show options for a dashboard parameter, then saved and closed the workbook, the parameter control would be blank upon reopening the workbook.
In views with a dimension on the pages shelf, mark labels visible on the first page would disappear when clicking through subsequent pages.
In workbooks with a large number of data sources, it was sometimes not possible to create or edit data source relationships for data blending.
Cloudera Impala table names were truncated to 18 characters in Tableau Desktop.
An internal error would occur when opening some workbooks, created in Tableau Desktop 9.0, in newer versions of Tableau Desktop.
When connected to SAP HANA, data ending in an escaped character could not be filtered correctly.
When using an Oracle-based data source, the DATEDIFF function would return different results depending on whether a live connection or extract was being used.
In some views with colored marks, and column or row grand totals shown, filtering out all mark values resulted in an error: "No such column [system:visual].[is_mark_total]"
New line and carriage return characters in SAP HANA data would cause errant queries to be generated, which resulted in non-functioning filters and other unexpected behavior.
In some non-English locales, you could not specify multiple percentiles using a semicolon separator when editing distribution bands in Tableau Desktop.
Thai text was too small in Tableau Desktop on Windows machines.
In some cases, when aggregate measures was turned off, modifying a date range filter in Tableau Desktop would result in a "type exception" error.
When editing map background images in some workbooks, an "Invalid predicate for image" error would occur.
Sheet captions tied to filtered field values would display 'All' even when no values were selected in the filter.
When using an extract data source based on Amazon Redshift, sum aggregated values for SMALLINT fields would display as positive and negative values in the range of -32,768 to 32,768, even though there were sum aggregated values greater than 32,768 and all of the original data in the field was positive.
Extra padding in IBM DB2 fixed-length string fields would get unexpectedly trimmed. This resulted in broken views when refreshing extracts of workbooks upgraded from prior versions of Tableau.
Moving column totals to the top or row totals to the left would incorrectly apply any background coloring that was applied to totals to dimension headers.
In workbooks that connect to a mix of extracted and live data sources, switching between data sources in the Data Source tab would often result in a missing password error.
In Tableau Desktop it was possible to set the minimum value of a range filter to be greater than the maximum value, resulting in a blank view.
Refreshing a statistical file based data source using the Tableau Online Sync Client would fail with an error: "There is no active connection to the data source"
Performance recordings created in Tableau 10.0 did not include "connecting to data source" events.
In workbooks using a Thai locale, changing a parameter data type from Date to String would add 543 years to the date displayed in the display format field.
The Tableau Online Sync Client would often crash when a user logged out in Windows.
Extra padding in Cloudera Impala string fields would get unexpectedly trimmed, which caused some queries to fail to return data.
When you tried to append data to an extract using custom SQL, an error would occur: "Tableau Data Engine Error: 48047: bad input"
Tableau Desktop sometimes crashed when displaying a progress dialog.
Using a top N filter with a context filter sometimes resulted in unexpected top values displayed.
Map tiles intermittently do not fully render in Tableau Desktop.
Tableau Desktop would often crash when you attempted to edit a large, complex data source. Same fix applies to issue 580719.