在 Tableau 9.0 版中，我们增加了很多功能，使产品更智能地处理您的操作。这些功能不仅包括帮助您将数据引入分析的自动数据准备，还包括更多分析功能，以及支持地理搜索的智能地图。我们还在性能方面作出了巨大投入。为了帮助您分享发现成果，协作处理数据，我们重写了 Tableau Server 和 Tableau Online。
下面的列表包含 Tableau 9.0 中修正的部分关键问题。
Quick filters in Polish listed the days of the week were listed alphabetically, rather than in calendar order.
When connected to an Oracle database, copying and pasting sheets between workbooks resulted in the "Oracle database error 1005: ORA-01005: null password given; logon denied" error.
When connecting to a data source, Tableau Desktop quit unexpectedly.
When connecting to an IBM Netezza database, using incorrect case for a database name resulted in the error:
For languages that Tableau is not localized into, the days of the week were sorted alphabetically, instead of the order of occurrence.
In Tableau Desktop, opening multiple instances of a workbook that is published to Tableau Server displayed incorrect workbook names, beginning with the second instance of the workbook.
For SQL Server, case sensitivity in the Describe Field dialog box was incorrect when custom SQL was used.
Editing a calculated member field in an MSAS data source caused the field to revert from a discrete field to a continuous field.
Using custom SQL to join tables in HIVE connections caused error: “FAILED: ParseException.”
Creating a full outer join using the generic ODBC connection to SQLite caused the error: “Unexpected error.” The full outer join type is no longer visible for general ODBC connections to SQLite.
Using stored procedures to connect to a SQL Server Parallel Data Warehouse (PDW) database caused the error: “The stored procedure "" does not return a result set, or the result set columns do not have names."
When adding or editing a join clause, the Join dialog box did not accept supported Asian characters.
On a Mac, when attempting to create an extract from an Excel file with a column name that included the Ñ character, the error "Tableau Data Engine Error 4: No such file or directory" would appear.
When you replaced a published data source, quick filters would disappear from workbook views.
When a URL parameter was used to change a table calculation sheet link action filter, a “LogicException” error would appear.
When connecting to Salesforce, joining tables resulted in the error: "Expression clause references too many tables".
Highlight actions failed in dashboards with tab names that contained multiple, sequential angle bracket characters or a single comma.
When connecting to Amazon EMR using Impala, creating an In/Out set, where the expectation is that all values would be in the Out set, instead results in a blank view. This is due to the way that Impala is optimizing the query.
Workbooks containing multiple nested table calculations that share common sub-calculations may render differently after upgrade to Tableau Desktop 8.3 or 9.0
If you encounter this issue, view the table calculation settings for the calculations that appear to render incorrectly. Verify that more than one table calculation is using an underlying calculation. Compare the settings between different instances of the calculation. Update any table calculations that reverted to default settings (for example, Table (Across)) to match calculations that retained the updated settings.
Clicking on a mark in a view caused the error: "tuple id cannot be resolved."
A global filter on a hierarchy field in Microsoft Analysis Services (MSAS) took changes made only in the most recent sheet included in the global filter. This sometimes caused sections of the view to display incorrectly.
When aliasing fields in the primary data source using the secondary data source, the "Edit Primary Aliases" command was not available. This command is now available if a dimension from the primary data source and another field from the secondary data source are being used in the view.