Known Issues

In a dashboard action, settings were not corrected saved when a worksheet name ended with a "\" character.

When opening a worksheet with multiple filters, an error would sometimes occur, "Tableau was unable to generate a query to perform this operation".

Calculations that use RAWSQL are not included in the GROUP BY clause when used as a Dimension with AWS Athena and Amazon Redshift.

Table names beginning with a number were displayed as a letter on the data source pane when connected to Google BigQuery.

The dialog, "Search for Data" would show blank after Tableau Desktop had been idle and the server session timeout had occurred.

When the Tableau Server idle timeout had expired, Tableau Desktop would not display a helpful error message to reconnect.

Error 6EA18A93 would sometimes occur when joining unioned tables.

Some workbooks created pre-Hyper would have poor performance in versions 2019.4 and 2020.1 when Context Filters were used.

In a logical join with a flat file, dragging measures from the CSV table to the view could cause a collation mismatch error.

The Schema list was not in order when connected to Amazon Redshift.

The auto field improvement to remove underscores would sometimes not work.

The error, "Type doesn't match expected type" would sometimes occur when changing a quick filter using Animations.

Starting with 2020.3 when a Set is placed on the filter shelf, it no longer affects the displayed list of values for all other Quick filters on the sheet.

In Tableau Prep Builder, sometimes bigint data was rounded incorrectly.

In the new feature for 2020.3.1 called Write to Database, if the database or schema does not have any tables (it is empty), navigating to the schema or database will not be possible. The schema needs to have at least one table created to be visible for selection.

Beginning with 2020.3.3, any existing rich discrete value filtering in the users flow or a new flow created in grouped steps, the annotation would not show on the profile card as expected to see the change or to edit it. The work around is the use the changes pane on the left or from the flow pane in the top to access the rich discrete value filtering annotation/change they had made.

A text file based workbook would sometimes cause the VizQL Server to spike and not release the CPU on Linux.

Tsm pending-changes list command would sometimes not provide output in the correct language starting in 2019.4.

Changing the data source refresh from incremental to full was not being implemented by Tableau Server.

Active Directory group synchronization would sometimes fail with "com.tableausoftware.domain.ldap.LdapSearchException:javax.naming.InvalidNameException" when special characters were used in the CN attribute.

Importing a configuration topology file with no "tdsservice" or "tdsnativeservice" would fail with an unclear error.

In Web Edit, a quick filter item with an empty (0 byte text) string value in a single value drop-down list, would not display.

Excessive 4096 Vizportal DataQualityService and WebClientDatasourceAppService errors would sometimes occur in Administrative views after an upgrade of Tableau Server.

Node became unlicensed when moving the license service to another node in multi-node server environment

Starting with 2020.3.0, the ‘nlp.exe’ or the Natural Language Process would take approximately 800 MB of memory where normally up to 250 MB of memory would be allocated. The increase may contribute to system instability on a new installation. A Tableau Server restart can resolve the issue until the product is updated.

On Windows Server 2012 R2 and Tableau Server 2020.3, degraded Viz load performance can occur when the New Relic Infrastructure Agent or other third-party monitoring software that uses WMI counters are running on the same node as Hyper and the Filestore.

During a new installation if the core key activated was not sufficient to handle the cores on the machine or VM, the installation would fail.

In some cases, the Tableau Resource Monitoring Tool would display 'Pending Deletion' for an extended period of time when an environment or server that is part of an environment was deleted by an administrator.

While executing a query in Impala the error "The table .... does not exist" may occur when refreshing view or setting filters.

Tableau Online admin views sometimes took a long time to load, or eventually timed out.

Connecting to Postgres via Tableau Online was successful, but error "syntax error at or near "ta"" was seen in the PostgreSQL logs.

When a scheduled extract refresh of a Quickbooks Online data source runs on Tableau Online, the following error occurs intermittently: "Intuit API Authentication error: Invalid authentication tokens. Login again and retry."

OAuth data sources published to Tableau Online such as Salesforce and Google Analytics would fail to refresh due to being routed to the Bridge client instead of running server-side.

When attempting to refresh an Amazon RedShift data source/workbook or run a subscription an error will occur, "Error trying to execute a query:[SQLState 22PO2] Error: invalid input syntax for integer".

Tableau Bridge was not always able to recover after a network failure

The edit button was missing for some WDC data sources.

Upgrading Bridge to 2020.3 would sometimes cause the client to freeze on "Loading Data sources".

The Bridge job process would sometimes be stuck when a network communication error was received getting the data source information from the Tableau Online server before updating the extract.

The Bridge client could repeatedly become disconnected where a restart was needed.