Google Chrome, Safari, Firefox

Google Chrome, Edge version 77+

logon

If you have multiple instances of Symphony running on the same server, and the paths to these instances differ only by the port number, users will be logged out automatically when they navigate from one instance to another instance and log on. Returning to the previous instance requires the user to log on again.

embedding, remote gateway

If you are using a reverse proxy, load balancer, or other network gateway that is remote (on a separate machine from Symphony), its IP address must be set in the advanced Symphony configuration setting Trusted Proxy IP Addresses.

embedding

Embedding Symphony on another webpage and applying the CSS display: none to the Symphony element or iframe will cause the browser to load the application without allowing its elements to render correctly. Symphony should be loaded when displayed, or you can use the alternative CSS visibility: hidden instead.

user interface

If your web browser has a plugin installed that adds a toolbar or occupies a portion of the browser real estate, this will affect the coordinate system of Symphony and can result in various UI issues.

deployment, Linux

Support for deploying and running Symphony directly on Amazon Linux 2 has been discontinued. Consider deploying using Docker containers or another Linux distribution instead.

deployment, cloud

Support for deploying to AWS Elastic Beanstalk by creating source bundles in the Deployment application is removed . Using Helm charts to deploy to Kubernetes is recommended instead for cloud deployments, and auto-scaling options are available.

data connectors

Discovery of scalar-valued functions and other non table-valued functions is not supported.

data connectors

If you edit an existing non-relational or non-OLAP data connector, make some changes to it (e.g., via Define Structure), and then perform a re-discovery, previously discovered structures will remain under the data connector in Structure), and then perform a re-discovery, previously discovered structures will remain under the data connector in Explore. If you then try to use these previous structures you may encounter connection errors.

data connectors, Access

Crosstab queries and linked tables are not discovered. The crosstab queries and linked tables are available if using manual queries. Another option is to create parameterless SELECT queries to the crosstab queries or linked tables (which will be discovered as views).

data connectors, ODBC, JDBC

A 64-bit version of the ODBC driver is needed to work with Symphony's 64-bit application pool.

Discovery of table and view definitions may not work for some ODBC and JDBC data sources. As a workaround, use data cubes based on manual SQL query.

Discovery of table relationships (e.g., foreign keys) may not work for some ODBC and JDBC data sources.

Stored procedures and functions are not discovered.

data connectors, SSAS

Executing certain valid MDX in SSAS 2012 fails with Internal error: An unexpected error occurred (file 'mdmisc.cpp', line 945, function 'MDBulkSetScaIter::Start'). Try installing SQL Server 2012 SP2 to fix this issue.

data connectors, SSAS, OLAP

The DRILLTHROUGH statement and other MDX queries that do not return a cellset are not supported.

data connectors, Microsoft Dynamics

Some entities in Microsoft Dynamics CRM/365 may not be supported, resulting in an error such as The 'RetrieveMultiple' method does not support entities of type... when using some tables.

data connectors, Access and DBF

The Microsoft driver used to connect to DBF files does not support FoxPro dBase 3.5. As an alternative, you can use an ODBC data connector to connect to the same files.

data connectors, IBM DB2

The IBM DB2 data provider may not be able to connect to certain versions of DB2 such as AS/400. ODBC or JDBC drivers can be used instead.

date mapping, OLAP

If the OLAP hierarchy being mapped has a data gap, using the option to manually set the start date will yield the wrong results. Specifically, years after the gap will be incorrectly mapped.

data cubes, warehouse storage

Strings longer than 900 characters (or unlimited length) cannot be indexed, which may cause performance issues when used in a data warehouse. Change the data type to a shorter set length to avoid this issue.

data cubes, warehouse storage

The floating point values NaN and Infinity are not supported and will cause the data warehouse build to fail.

data cubes, in-memory storage

In-memory storage relies on the double numeric type, which may approximate rather than guarantee accuracy when displaying a large number of significant digits (e.g., many digits to the right of the decimal point).

data cubes, hierarchies, in-memory storage

Previewing a hierarchy in the Hierarchy editor may be slow if the underlying data cube is using in-memory storage. You may get better performance by using warehouse storage for the data cube instead of in-memory.

metric sets, period over period

Period over period cannot be added to an existing period over period measure.

metric sets, natural language assistant, Python

When Symphony is installed on some instances of Windows 10 version 2004 or higher, the Python machine learning engine may fail to start when users try to use the metric set assistant due to a Windows issue combined with NumPy version 1.19.4. Please note that only Windows Server is supported for production installations of Symphony.

You can avoid the error by installing NumPy version 1.19.5 or later from the command line on the server (e.g., pip install numpy --upgrade) or with an update to Windows.

metric sets, formatting

Standard numeric formatting is provided by .NET and will round midpoint values to the nearest even number (e.g., 1.65 rounded to 1.6 because 6 is even), which may not align with expectations. Custom numeric formats round away from zero instead (e.g., custom format 0,0.0 applied to 1.65 results in 1.7). You can also round midpoint values away from zero if needed when calling Math.Round in a formula or calculated element, for example: Math.Round($Measure$, 1, MidpointRounding.AwayFromZero)

metric sets, Oracle

Oracle Character Large Object (CLOB) data used with a parameter compares only the first 2000 characters. This may cause extra rows to be returned.

metric sets, OLAP

When using an OLAP cube in cases where the application is performing the aggregations rather than the cube, you may need to change the aggregator from Provider to a specific alternative to avoid missing values. For example: when a formula is added based on the measure and the Collapse All option is used on a hierarchy.

metric sets, dynamic hierarchies, dynamic measures

Dynamic hierarchies and measures are not meant to replace editing a metric set for performing analysis, and have more limited filtering and sorting options.

metric sets, dynamic hierarchies, period over period

Period over period cannot be used with a dynamic hierarchy selected as the time hierarchy.

metric sets, dynamic hierarchies

The No Selection token is provided for dynamic measures only and is not supported for dynamic hierarchies.

formula metric sets, dynamic hierarchies

If a dynamic hierarchy is used in a formula visualization, its level should be set the same as in the original metric set.

filtering

The Single Number filter should only be used to filter integer data values. To filter decimal type data, use the Range Number filter instead.

reports, scorecards, small multiples

Reports, scorecards, and small multiple views do not support metric sets with data retrieval format set to Raw Data. These are metric sets created when you drag a database table directly onto the canvas.

reports, scorecards, small multiples

Reports, scorecards, and small multiple views do not support the Frame component.

reports, scorecards, small multiples

Zooming and scrolling (scrollbars) are not supported for charts appearing in a scorecard, report, or small multiple.

reports

It is recommended not to create a single item in a report that spans more than one page. When viewing reports within the application, a single row expanding to fill more than one page may be displayed contiguously without separation from page margins, headers, or footers. Exporting to PDF with an item spanning more than one page may cause other unexpected results.

dashboards

Positioning items outside of the canvas and moving them in through script is not supported. Consider using layers or the hidden property instead.

dashboards, data input

When setting up a data input interaction, writing to a data input transform in an inherited data cube is not supported and will not be available as a selection.

PDF export, image export

A state indicator that displays an image may appear blurry when exported to a PDF or image file. This is due to a known issue with the Chrome browser relied on for all exports. Using a data image instead can avoid the problem.

Excel export, OLAP

Exporting to Excel with the measure format set to OLAP will not include the number formatting.

Excel export, table

When exporting columns using Group Text, the first grouped column's group header settings are applied in Excel.

Excel export, table

Table columns using a color rule with distribution set to Round Interval or Round Continuous are treated as Equal Interval and Continuous, respectively, when applied to the spreadsheet's cells when exporting to Excel.

Excel export, table

When multiple states from different state groups apply to the same cell, only one state may be applied in the export file.

re-visualizing

When re-visualizing a data visualization from one kind to another, only the underlying metric set is maintained and not any visualization-specific properties. This is because a table does not support chart-specific styles, for example. To get the original visualization and its styling back when viewing, reopen the file or use the Revert To Original re-visualize option. When editing, use undo in the toolbar, or first check in the file, use the Update button on the Data Analysis Panel, or save a style.