Known issues and limitations for SPSS Modeler

The following known issues and limitations apply to SPSS Modeler.

For more information about troubleshooting other issues in SPSS Modeler, see Troubleshooting SPSS Modeler.

Known Issues

upgradeStatus gets stuck in InProgress status after upgrading IBM® Software Hub

Applies to: 5.4.0

When you upgrade SPSS Modeler for IBM Software Hub Version 5.4.x, the upgrade completes successfully, but the upgradeStatus might get stuck in InProgress status.

This status inconsistency occurs because the set_status playbook didn't trigger when the upgrade completes.

Workaround

You can ignore the upgradeStatus if the status for the custom resource (CR) is Completed.

Component           CR-kind              CR-name                           Namespace    Status      Version    Creationtimestamp     Reconciled-version       Operator-info
spss                Spss                 spss-sample                       base       Completed  10.0.0      2025-12-12T13:33:46Z  10.0.0                    10.0.0+20240906.223839.13
Rows with no documents don't display correctly in Categories tab

Applies to: 5.4.0

This display issue happens in the Categories tab of the Text Analytics Workbench. If a category does not have any matches in a record or document, the row sometimes displays a refresh icon instead of a 0.

Figure 1. Categories tab showing an icon instead of a number
Image of the Categories tab showing an icon a refresh icon instead of a number 0.

Aside from the display issue, the Categories tab functions normally. It does not need to refresh the row.

Workaround

Treat any rows with the refresh icon as having returned no matches in records or documents.

Data assets from Teradata connection

Applies to: 5.4.0

For Teradata instances, if the Transaction Semantics column shows Teradata, then SET tables are created by default. If the Transaction Semantics column shows ANSI, then MULTISET tables are created by default.

SET tables prevent duplicates. MULTISET tables allow duplicates.

If your system creates SET tables, you might encounter the following issues:
Duplicate Row error
For some Teradata instances, if the rows to be appended are already in the table, a Duplicate Row error might occur during the Append action.
Hang issue
You might notice some hang issues. SET tables might have a performance impact.

Limitations

Downtime for SPSS Modeler during upgrade to IBM Software Hub Version 5.4

During the upgrade to IBM Software Hub Version 5.4, jobs for SPSS Modeler are executed, but all builds fail.

This behavior is expected. The upgrade waits for the existing session to terminate, and then it starts to apply the new runtime definition.

Workaround

Wait until the upgrade completes and try again.

Languages supported by Text Analytics

SPSS Modeler Text Analytics supports the following languages:

  • Chinese (Simplified and Traditional)
  • Dutch
  • English
  • French
  • German
  • Italian
  • Japanese
  • Portuguese
  • Spanish

The Text Mining node can extract text data written in both Simplified Chinese and Traditional Chinese. However, SPSS Modeler currently provides templates and text analysis packages only for Simplified Chinese.

No SQL label for Data Asset Export node when you use watsonx.data™ Presto

When you use data from the watsonx.data Presto connection and SQL pushback is enabled, the Data Asset Export node does use the SQL pushback feature. As a result, Data Asset Export nodes do not have the SQL label in the user interface.

Data streaming for Google BigQuery with SQL pushback

If SQL pushback is enabled, data streaming is not used to insert data when you select the Streaming Insert option for Write mode in a Data Asset Export node. A standard insert occurs instead. If you want to use data streaming with Google BigQuery, you must disable SQL pushback.

Special characters for Google BigQuery with SQL pushback
Google BigQuery does not support some special characters (like $ and *) in column names, and SQL validation fails for queries that contain these characters.

To avoid errors, SPSS Modeler automatically disables SQL pushback for any node where a special character appears in a column name. SQL pushback is also disabled for all downstream nodes in the flow.

Workaround

To use SQL pushback with your data, you can add a Filter node to your flow to filter out columns or change their names. You must rename or filter the model output fields that contain $.

  • To change the name of a column, edit it in the Output field column.
  • To filter out a column from your data, click the arrow in the Filter column.

For a list of supported special characters, see Specifying a schema in the Google BigQuery documentation.

Timestamp data measured in microseconds

If you have timestamp data that is measured in microseconds, you can use the more precise data in your flow. However, you can import data that is measured in microseconds only from connectors that support SQL pushback. For more information about which connectors support SQL pushback, see the documentation:

Merge node and Unicode characters
The Merge node treats the following similar Japanese characters as the same character.
Figure 2. Japanese characters
Similar Japanese characters treated as the same by the Merge node