Loginom 6.4.3 release notes

In this update, file storage was upgraded, performance improvements related to operation of processor and memory in the multithreading mode were made, and some fixes were performed

Fixes

Loginom Server: when several sessions with interconnected packages were simultaneously completed, the following error could occur: "Index value of Items parameter is out of range" that was recorded in the server log (#8470).

Loginom Server: when using Apache web server with the configuration created by default, connection via Web Socket Proxy failed. The error was caused by the fact that WS_PROXY_URI parameter in the created env.conf file was set as ws/. The correct paramater value is app/ws/ (#8544).

SOAP request: if WSDL schema contained two items with the same names one of which was the root one, the following error occurred while request sending: "Input XSD schema does not comply with the node settings!" (#6179).

Quick view: in the Quick view detailing, values of the items with "Variable" type were displayed as  (#8630).

Node Execution: if in "Node Execution" settings the node for which the execution order had been earlier set as the selected one in such a way that "Node Execution" node was to be performed earlier than the selected node, the following error occurred: "Next move forbidden. Circular links between nodes are not allowed". The same error could occur while opening the package containing the same node configuration. Such situation does not represent a circular reference as only node configuration is used for "Node Execution" as opposed to "Reference Node", and data that differ from the data that is used in the base node can be supplied to the input (#8614).

Date and Time: when transforming dates to the string format, there were no leading zeros for year week, quarter week, year day and quarter day. The default month day format is made with leading zero. The default Date and Time format was Fixed ("2021-01-30""23:59:59""01:01:01") (#8595).

Import from Deductor Warehouse: if the package with configured import from Deductor Warehouse is saved and repeatedly opened, while attempting to execute the import node, the following error occurred: "Deductor Warehouse: Data Warehouse object not specified" (#8523).

Import from Deductor Warehouse: if one and the same Dimension was included into the Process several times, and the caption was changed while designing, these Dimensions had the same names in the import wizard in the parameters tree, but when transferring such Dimensions to the Fields area, their caption texts were displayed correctly (#8556).

Import from PostgreSQL: when one and the same parameter was used more than once in the request, the following error occurred: "Invalid parameters specified" (#8587).

Binary classification assessment: when copying data from error matrix to the clipboard to the values in Predicted column, extra % symbols were added (#8638).

Coarse сlasses: when moving in the wizard back from the binning configuration page and then backwards, if the settings that had influence on class count had been changed, the previous class count was displayed on the wizard pages. It could also cause the following errors: "Next move forbidden Failed to apply fine classes for binning for %s column. There are too many unique values" or "Value 0 of "Index" parameter is out of [-1; -1] range" when switching the wizard pages (#8282, #8582, #8584, #8668, #8700).

Coarse сlasses: while merging classes with continuous data type, the upper and lower bounds of the updated classes were incorrectly defined (#8577).

Coarse сlasses: "Merge with previous" and "Merge with next" buttons were always active in the binning editing mode. If one of the buttons was pressed upon classes merging, classes were merged, in spite of the fact that none of the classes had been selected. Now, when enabling the editing mode, the selected class is clearly displayed, and "Merge with previous" and "Merge with next" buttons become inactive if the action cannot be performed (#8578).

Cube: if after import from Deductor Warehouse or LGD file in which Group default usage types were set for fields, the Cube visualizer was created, and the displayed measures were configured for the fields with Group default usage types. Then, the calculated measure based on the measures selected earlier for display was added. Afterwards, the measures used for calculation of the calcualted measures were concealed, and after package saving and repeated opening, the calculated measures were not calculated (#5803).

Cube: while package saving, order of the fields transferred to the Cube filtering area was not saved (#8490).

Cube: if the number of dimensions in columns was higher than the number of columns in the cross table, some part of dimension captions was not displayed when exporting. Such situation could arise if column totals display was disabled (#8639).

Cube, Supernode: if the Cube visualizer with the sorting configured by one of the columns was configured in the base Supernode, while attempting to configure the derived Supernode, the following error occurred: "__baseName attribute does not exist" (#8536).

Cube, Supernode: if sorting was configured for the node containing the configured Cube visualizer in the derived Supernode, and then dimensions in the Cube were added for the base Supernode node, the configured visualizer was marked as the invalid one upon the node configuration update (#8528).

Batch processing: if the package file to which the processed package referred was deleted or renamed during the package batch processing, upon the main package processing, the following message was recorded to the server log: "Error accessing file: ...lgp", and the processed package remained in the server memory, whereas the deleted or renamed package was still displayed in the list of common packages in the Session Manager (#8572).

Connections to FireBird DB, Interbase: upon database server connection loss, connection was not restored on repeated connection activation (#8615).

Supernode: if the node connected to the output Supernode port was activated, and the Supernode itself was not activated, and then the package containing the Supernode was saved and closed, on repeated package opening and while attempting to activate the Supernode, the following error occurred: "Port not configured" (#8538).

Supernode: if the base Supernode was changed upon creation of the derived component and derived node (for example, one more node was added to the supernode), while attempting to activate the derived node, the following error occurred: "Internal Supernode nodes locked". The error occurred in version 6.4.2 (#8629).

Table: the value of Base interval date parameter was not defined and saved in the filter configuration window in the Table visualizer for the fields of the Date/Time type, when selecting first / not first or last / not last conditions (#8648).

Table to tree: if the automatic mapping of the input and output fields failed, the following error occurred upon the node configuration: "Table to tree: The output node ... has not been mapped to the input node". It was required to go to the output port settings to perform the fixes. Now, if such situation arises during the node configuration process, the following page is opened: "Adjustment of Data Trees Mapping" (#8659).

Reference node: when saving the package containing the Reference Node, the following error could occur: "..." node is locked" if the node that used the data of the output port of the source node for which the Reference Node had been configured was executed in the parallel session (#8680).

File storage: if the new directories became available to a user during his current session, while attempting to go to such directory, the following error occurred: "Cannot read property 'expandAsync' of undefined". For example, such situation could arise if another user gave access to the user, the rights of access to the directories via the Administration bar were changed, or the new directory was created via the file saving dialog (#5347, #6032, #6423, #8022).

File storage: if the internal directory was moved one level up or down from the directory, while attempting to open the first directory, the following error occurred: "System Error. Code: 3. System failed to find the specified path" (#6674).

File storage: if one of the directories containing the copied directory was renamed after directory copying, when pasting, the following error occurred: "System Error. Code: 3. System failed to find the specified path" (#7082).

File storage: if the file was uploaded and then deleted, it was impossible to upload it again with the same name before the application session termination. The error occurred in the browsers based on Chromium, in Internet Explorer and Safari (#7108, #8058, #8499).

File storage: when moving from the current directory to the parent one, ".." was selected but not the quitted directory. Incorrect focusing could cause incorrect hotkey operation of copy-paste functionality (#8023, #8516).

File storage: while attempting to open the package using the symbolic link to the file physically located in the network folder, the following error occurred: "System failed to find the specified path". It was connected with failure to define the file modification date (#8452).

File storage: the folder deleted by means of "Cut-Paste" functionality was not deleted from the directory tree. While attempting to move to it in the directory tree, the following error occurred: "System Error". Code: 3. System failed to find the specified path" (#8518).

File storage: if it was required to go to the file storage folder that had been previously opened in another tab, and then try to close the tab with the previously opened folder by pressing X (cross button icon), the following error occurred: "Cannot read property 'ActiveLoadersCount' of undefined" or "this.FFileSenderManager is undefined" (#8529).

Row filter: when repeatedly entering the filtering wizard, the tile caption with the filter by the time lapse was not fully displayed in case of the long text. For example, if "the first day from the current date" was selected (#8633).

Row filter: it was impossible to set two and more conditions by the time lapse (first/not first or last/not last) (#8634).

Loop: on activation of the parallel Loop containing the nested Supernode and Reference Node referring to the node located in the Supernode, the following error could occur: "Loop" node is locked". The higher number of iterations was selected in the Loop settings, the higher risk of error occurrence was identified (#8531).

Export to PostgreSQL: while attempting to export strings to xml or json fields, the following error occurred: "column "..." has type ..., and expression - character varying". If the error has been identified before, it is required to go to the export configuration wizard once again after update (#8585).

Platform core: if the Supernode contained the Reference Node referring to another node inside this supernode, and the execution order of the Reference Node was connected with the input and output port of the Supernode that contained it, the Supernode activation process could be frozen, and it was impossible to cancel it. If the Supernode was executed in the Loop, it increased the risk of error occurrence(#8514).

UI/UX Improvements

Import from Deductor Warehouse: while attempting to execute unconfigured import node, uninformative error occurred. Now, the warning that there is no selected data warehouse object is displayed (#8515).

Configure ports: "Socket is not configured" and "No reference to socket specified" error messages are replaced with the following more informative messages: "...: Port ... not configured" and "...: Link to the required port ... does not exist" specifying caption of the port in which the error occurred (#6000).

Ease of Use

Python:builtin_pandas_utils.to_data_frame function was improved to provide the possibility of pandas.DataFrame creation by the integer and logical columns containing null data. If it is required to process the null values related to the integer and logical types when filling  numpy arrays, the real type is used, and null values are transformed to NaN (np.nan). Previously, while attempting to transfer the table containing the integer or logical column with null values to builtin_pandas_utils.to_data_frame function, the following error occurred: "...int() argument must be a string, a bytes-like object or a number, not 'NoneType' " (#8522).

Import from Deductor Warehouse: automatic initialization "Default usage types" based on the data warehouse objects types is disabled for the output columns. Previously, "Default usage types" equal to Value, and for Dimensions - Group were set for Measures . The User can set "Default usage types" for the output columns at his discretion. The previous behaviour could cause the errors also in the Cube visualizer (#8506).

Coarse сlasses: view with rounding of values of the coarse class bounds is disabled in the wizard. Previously, their value was displayed with rounding up to two decimal places in case of the manual editing of bounds (#8635).

Logging: node GUID (unique identifier) was added to the log messages for the events relating to the node execution start and end (#8667).

File storage: when executing "Copy-Paste" functionality, move operation is now used instead of the source folder copying and pasting. Previously, "Cut-Paste" operation could take much time, and when copying 2 GB directories, it was possible to cut the subdirectory of the copied directory in the next tab while copying (#6031, #6366).

File storage: the speed of directories creation is optimised for the case when there are many directories in the current directory (several dozens or hundreds of nested directories) (#8196).

File storage: the actions that could not be performed became inactive in the file storage root: Load, New Folder, Download, Rename, Copy, Paste, Cut, Delete (#8542).

#loginom#release notes#6.4#6.4.3

See also

Loginom 6.5.2 release notes
The error when executing "Python" node in the loop with the group processing was corrected. Logging when working with several packages at once in parallel was improved. Some errors occurring...
Loginom 6.5.1 release notes
Quick view interface was improved. The memory leak error was corrected in Python. Version 3.10 is now supported. Errors in “Coarse classes”, “Logical regression” and other handlers were...
Loginom 6.5.0 release notes
Major Update of Loginom Analytical Platform Quick view of the calculation results functionality was implemented for selection of the activated node. It became possible to cancel the request...