The Python operation with numexpr module was corrected in the code execution mode inside the process, errors occurring while executing requests to databases were corrected, regression bugs in several handlers were eliminated. It became possible to activate JavaScript and Python nodes data sources ports of which are connected to the inactive Condition node branches.
Python: The Python operation with numexpr module was corrected in the Python code execution mode inside the process. Previously, when using pandas and numexpr package, start of the Python node caused the following error: "Offset exceed JPM opcode limit". (#9403).
Cube visualizer: In some cases, identifiers of the Cube headers were transferred from the server in the reduced form. The required corrections were made to avoid possible errors while displaying the Cube cross table headers (#8708).
Import from Excel: After closing the node wizard in import from Excel, Object name parameter setting set by the variable was not saved. That's why it was not possible to import from several Excel workbook sheets in the loop(#9266).
Import from database: The input variables were not activated after the first start of the SQL request results preview in the import from DB wizard. In the result, if the request contained variables as parameter or macro lookup, a user received the following error message: "Error when opening SQL: SELECT from ... near "from": syntax error" (#9353).
Coarse сlasses: The set Configured "event" value was not saved in the node wizard for the output discrete field related to the variable type. After saving and repeated wizard opening Configured "event" value was reset to null (#9356).
Cross table: Overflow checking was added for the number of resulting columns.
PostgreSQL connection: The following error was corrected: "column "value" is of type json but expression is of type character varying". It occurred when trying to export the first row to xml or json fields in PostgreSQL DB (#9247).
SQLite connection: The requests that contained LIKE and did not end with % character were not operable in SQLite with the default connection settings ("Case-insensitive" LIKE operation mode). The required corrections were introduced (#9361).
Row filter: Variables in the Row filter handler became inoperable in the workflows created up to 6.5.0 version, if the field and variable types were incompatible (#9284).
Row filter: The preview did not change data in the node wizard when changing the variable used in the filter condition (#9423).
JavaScript, Python: It became possible to activate JavaScript and Python nodes, the data sources ports of which are connected to the inactive branches of the Condition node. If there are input ports connections, at least, one input port must be connected to the active branch. The input variables port must be connected to the active branch or there must be no connection (#9367).
It is required to pay special attention to the following facts connected with the introduced changes:
SQLite: Sqlite library loading was optimised.
Import from text file: Processing of Date/Time type values was optimised. In the result, import from text file became by 1.5 times faster.
