Loginom 6.3.2 release notes

Work with packages distinguished by big size or complex structure was optimized. Several important user identifications were added. This release fixes some important bugs.


  1. ARIMAX: when model training in the auto fit mode, not all exceptional situations were correctly processed (#6809).
  2. REST request: if Content-Type was not specified in the reply, there was an attempt to parse the reply as XML.  If the reply body was not XML, a user received the following message: “Error while XML document parsing: …: the server reply was not XML file” (#7321).
  3. Date and time: formats of the date to %t and %m row transformation were not saved (#7559, #7569).
  4. Chart: time caption value in tips differed from the real data by hour zone value (datetime + UTC) (#7563). 
  5. Import from Databases: method of receipt of table field descriptions for DBMS Oracle changed from the SELECT request formation to the request form the metadata content. Previously, an error could occur in the wizard of import from database for Oracle, for example, there were no sufficient rights for receipt of table field description using the old method (#7288).
  6. Property inspector: when opening the node containing significant amount of subelements, the following error occurred: “Cannot read property 'down' of null” (#7572). 
  7. Calculator: RegExDomain calculator function description was corrected (#7019).
  8. Calculator:  CumulativeSum function was incorrectly calculated if the field calculated in the same Calculator node was used as an argument, and request to the values of such field was random (#7575).   
  9. Calculator: if the string type data of the output fields were requested in several parallel streams, errors could occur, including the “frozen” execution of workflow without any messages about errors (#7596).  
  10. Binning: when using the Tile binning method with definite number of resulting bins, an error could occur (#6995).   
  11. Binning: if one of the following fields Lower bound, Upper bound, Deviation from the lower bound or Deviation from the upper bound had the Variable type of data, the following error message occurred, when downloading External bins at the stage of their configuration: “Next move forbidden. “Active” usage type is not valid for the configurable column…” (#7327).
  12. Binning: when using the External ranges method, the value of the Upper bound open flag was ignored, and the value of the Lower bound open flag had impact not only on the upper but also on the lower bound (#7486). 
  13. Coarse classes: in the Class parameters output port values of the lower and upper bounds of the first and last coarse classes were incorrectly displayed as equal to “0” (#6801).
  14. Coarse classes: significance of the null class in the first output port for continuous data had incorrect value equal to “-1” (#7219).
  15. Correlation analysis: calculation of Kendall's Tau-b correlation coefficient was corrected. Possibility of the partial data range overlapping with repeated elements was not considered, as well as the presence of fields with null values (#4743).
  16. Correlation analysis: calculations of extremum of cross-correlation function and Spearman's rank correlation were corrected in the case of fields containing null data (#7407, #7443).
  17. Cube: when changing the chart window width on X axis between the set values of the dimensions of the Date/Time type, fractional numbers could be incorrectly displayed (#6287).
  18. Cube: it is possible to add the existing but unused measure once again in the measure configuration window using menu of the “Add measure…” button that caused errors (#7506).
  19. Cube: when configuring the data display format, the new value of the Horizontal Alignment format and Negative in red display variant was not applied (#7587).
  20. Linear regression, Logistic regression: calculation of some linear regression parameters was corrected, and F-test (Fisher's ratio test) procedure performance was changed. Availability of probability ratio of models inclusion and exclusion according to the chosen method of parameter selection was corrected in the linear regression wizard. Calculation of confidence range was added if corresponding option was selected in the course of node configuration. When calculating Akaike, Bayes and Hannan-Quinn criteria, only controlled parameters are currently used that has been recently recommended in the community. It allows for comparison of assessed models based on the samples of different volume (#7423, #7622).
  21. Adjustment of columns mapping: when adjusting mapping in the tabular mode, the following error could occur “Cannot read property 'set' of undefined”, while attempting to replace blank link with blank link for one of the fields (to select the blank cell in the first mapping table column) (#5771). 
  22. Adjustment of columns mapping: no sorting icon appeared in the caption of the column containing field captions (#6261).
  23. Neural network: cancellation of training performed for the large data volume could last for a long period of time, and the software was blocked till accomplishment of the training cancellation process (#6216, #6675).
  24. Navigation toolbar: in the case of several opened packages the search was available in the navigation toolbar window only for the first opened package, and no search was possible in other packages (#7474).
  25. Supernode: when deactivating the nodes connected to the Supernode inputs, the input Supernode node was not deactivated if the Supernode was not active. When the output Supernode node deactivating, the nodes connected to the Supernode outputs were not deactivated, and the Supernode was not deactivated (#3319). 
  26. Supernode: when the package opening during the node configuration loading, errors could occur. For this reason, the loaded node was nonworking. It occurred in the case when the downloaded package contained workflows with complex dependence between nodes. Example of the complex dependence: the Package contains two Modules. The first one contains Supernode #1 that includes the Node inherited from the Derived Component, the base Node of which is in the second Module. Besides, there is Supernode #2 inherited from Supernode #1 in the first Module, and changes have been introduced into the settings of one of the internal Nodes of Supernode # 2 (#4917). 
  27. Supernode: if during activation of one Node inside the Supernode and before the activation termination the other Node also located inside the Supernode was activated, then while attempting to open settings of one of the Nodes, the following error could occur “When switching to “…” error occurred. The “Supernode” node is locked” (#5854).
  28. Supernode: when changing input ports of the base Supernode, the Supernode derived from it lost earlier configured visualizers during consequent synchronization with the base Supernode (#6855).   
  29. Web service publication: in the case of change of the package to which earlier published package refers, there was a situation when outdated version of the base package was called from the published package. Also the following error could occur in this situation: “The local component already exists in {...-..-..-...}  identifier list” (#5781, #7526). 
  30. Table: when copying many rows (several thousands) to clipboard, a user will receive the following notification: “Failed to copy selected cells to clipboard. Too many rows selected”. The following errors could occur in this case: “Cannot read property 'down' of null” or “Cannot read property 'data' of undefined” (#7551, #7584).
  31. Reference Node: if when copying-pasting restoring of the Reference Node and the source node link failed, a user will receive the following warning: “Failed to restore link from the “Reference Node” node. “...” source node is not available at the required destination”. The Reference Node has been earlier added in the inoperable state. But there were no error messages (#7541). 
  32. Filtration: sometimes date and time setup panel worked incorrectly that caused impossibility to set time using sliders (#5987).
  33. Loop: when terminating execution in the parallel Supernode Loop containing nested nodes, an error could occur that resulted in failure in attempting to finish the application operation in a correct manner (#7499).
  34. Loop: when activating the cloned trained Loop, containing the nested Loop, the following error occurred: “Port not configured” (#7578).
  35. User interface elements:  when deleting a node, using the Navigation toolbar, no window with the deletion confirmation appeared, and it was possible to delete the configured node (#6106).
  36. Core framework: if a package with the base node was not available at the moment when the package containing the nodes depending on the base one was being saved, dependence was not restored when package opening subsequently. It caused null data set in the case of subsequent execution of the Reference Node of the dependent package on its output port (#6397).

Performance improvements

Cross table: memory consumption was optimized (#7498).

Web service publication: memory consumption for publication of large packages was optimized (#3445).

Core framework: time of saving of the workflows containing the derived nodes configuration of which includes large amount of unique values was optimized (#7602). 

Core framework: time of loading of the package containing the nodes with large amount (thousands) of input and/or output fields was optimized (#7195). 

UI/UX improvements 

Core framework: a new user notification was added in attempting to open the package saved in the higher software version (#7373).

#release notes#6.3#6.3.2#loginom

See also

Genetic algorithms as a mathematical apparatus
Genetic algorithms aim to solve optimization and modeling tasks by iterating, variation and combining target parameters with the help of mechanisms similar to biological evolution.
Scalable CLOPE algorithm for clustering categorical data
Splitting of categorical and transactional data sets into groups with similar attributes into large databases is the most important task of data mining. In most cases, traditional clustering...
Clustering algorithms on Data Mining
This article is an attempt to systematize and give a holistic view of the latest achievements in the development of effective approaches to data clustering. The purpose of this article was not...