Megaladata 7.3.1—Release Notes
Errors Fixed
Megaladata Integrator
Previously, the output tree of a published supernode allowed required nodes only; optional nodes were not supported. If a required container was missing from the output tree, the Megaladata Integrator would trigger an error and prevent the package from returning results. This has been resolved. All output tree nodes are now optional by default, with the exception of the root node. (#11958)
JavaScript
In the JavaScript node configuration wizard, the items in the UsageType enumeration incorrectly included a ut prefix. This has been corrected to display the proper enumeration names. (#11612)
Python
In Preview, environment variables were failing to pass to the Python process, even when the Pass node environment variables option was set to true in the Python component parameters (located in the Administration section). This prevented scripts like python_run.sh from executing correctly during preview. This issue has been resolved. (#11955)
Fixed a rare race condition where a "Try again" error ('Unable to open file ".../outputtables.metadata": Try again') occurred sporadically. This happened when a Python node attempted to read the outputtables.metadata file while running in parallel with other processes. (#11893)
SQL Script
Fixed an error in the SQL Script node that occurred when an input table was empty. If a query was configured to use input values as parameters or macro substitutions, attempting to activate the node would trigger a "Row number 0 is out of [0, -1] range" error. The node now handles empty input tables correctly. (#11960)
Cube visualizer
Fixed a regression error (first appearing in version 7.0.2) that triggered the message "Index value of "Items" parameter is out of range" in the Cube visualizer. This occurred when a dimension used the In original order sorting option and the cross table had been sorted at least once; if the visualizer was then run on a new dataset with more columns and more unique values in the measured column, the process would fail. This issue is now resolved. (#11979)
Grouping and Table to Variables
Fixed a bug where opening a version 7.3.0 package containing a Grouping node would trigger a "List index (9) out of bounds" error. This issue was caused by the node's autogenerated caption logic and has been resolved. (#11970)
Excel file import
A rare error "Invalid row/column index" was fixed. It would appear when reading a non-standard Excel file from a billing system (files not created in MS Excel) (#11947).
Calculator
Fixed a regression error in the Calculator (Variables) node where the IsNull function returned an incorrect value when processing intermediate expressions of any type other than string. This issue, which first appeared in version 7.3.0, has been resolved. (#11942)
The behavior of the EncodeDate function in the Calculator node has been updated. If an argument value is outside the allowed range, the function now logs an error and returns Null. (#7104)
Field and variable mapping
Fixed a bug in the Column/variable mapping settings windows where the Delete selected context menu command became unavailable in Links interface mode. This occurred if multiple items were selected and one was deleted before attempting to use the context menu on the remaining selection. The command now remains available as expected. (#11897).
When cloning a variable, an uneditable Default usage type field (set to Unspecified) would appear erroneously in place of a configurable Usage type field. Now fixed: the preset usage type is Used, and the field is editable. (#11953)
Normalization
On the Normalization settings wizard page for data mining nodes, dates in the Range control section incorrectly appeared as numbers. This has been corrected; dates for both the controlled and source ranges now display in the Date/Time format. (#5356)
Tree Union
Fixed a bug where a Tree Union node failed to activate if located in an inactive branch after a Condition node. This configuration previously required encapsulating the nodes within a supernode to function; the Tree Union node now activates correctly in all branches. (#11981)
Processes pane
Modified the logic of calculating the remaining time for processes which have subprocesses. Previously, the remaining time of a process could be shown as less than that of its subprocess. Now fixed (#11936).
MS SQL connection
When searching for ODBC drivers under Linux, the system failed to find the FreeTDS or MSODBCSQL drivers if one of them had been registered more than once. Now fixed.
PostgreSQL connection
Fixed a bug in Database export nodes using a PostgreSQL connection. If you entered a nonexistent table name on the Export to database wizard page and proceeded to the next step. returning to that page to select an existing table would trigger a transaction error: "Current transaction is aborted, commands ignored until end of transaction block". This issue has been resolved. (#11312)
Added support for export to the interval type fields. Now, interval fields are recognized as string, so the exported values should be compatible. If automapping was disabled when configuring the Database export node, you need to go through all wizard steps again to enable export to interval type fields.
Fixed a regression error "column "interval_column_name" is of type interval but expression is of type character varying" which occurred when exporting to interval fields. (#11881)
Supernode
When you tried to change the type of a previously added port (e.g, from table to variables), the port's name would change after saving and the port would become optional, but the type wouldn't change. This regression error first appeared in version 7.3.0. Now fixed (#11971).
Application
Fixed a critical error (introduced in version 7.3.0) that caused the application to freeze during Database export, most notably when using ClickHouse. While the application was in this state, any authentication attempt on the Login page would fail with the error: "Failed to connect to server. Error: Connection closed by remote host." This issue has been resolved. (#11996)
Fixed a bug where package variable description would disappear after saving and reopening a package. These descriptions are now correctly preserved. (#12006)
Sorting
When sorting in parallel, some data didn't get sorted sometimes. This error is hard to reproduce during standard application use. Fixed. (#11962)
Export to database
Fixed an "Access violation" error that occurred when exporting to an MS SQL database on Linux if the connection was configured with the Do not use DBMS client option enabled. This error consistently appeared if the export process was canceled. (#11930)
Fixed a bug in Database export nodes using Oracle, MySQL, or ODBC connections where entering a single dot (.) in the Table name field incorrectly loaded a list of all fields from every table in the database. The node now performs proper validation for compound names: it returns the error "No table name specified" if the table name is missing, or "No schema name specified" if the database is specified but the schema is omitted. (#11997)
Previously, when working with MS SQL through ODBC, you couldn't set up export to a temporary table. (The temporary table's fields didn't appear in the Database export node's wizard.) This behavior is fixed now. (#11937)
Temporary tables are now supported even in single-session mode.
Other improvements
The Description field in the variable setup wizard now supports multiline hints. The Description column now displays multiline text more conveniently. (#11933)
See also