Grouping
Last updated
Was this helpful?
Last updated
Was this helpful?
In the tab Grouping the loaded data set can be filtered and grouped according to properties of the executions. The groups formed here are the basis for further analysis.
Groups are a central topic in the Report Analyzer. They are used to compare data sets according to arbitrarily defined criteria. These can be different product types, but also different users or characteristics of the test item or the test environment. In MSA analysis, groups are given additional significance in certain cases.
The prerequisite for this is that there is also data that allows grouping. Please refer to the section on optimizing report data (adding header information)!
For all created groups, a statistical analysis is performed for each measurement. Some analysis tools allow to compare the results of the groups.
Even if you have not defined any special properties, there are still many standard fields that support grouping and filtering (e.g. SeriesNr, BatchNr, SocketNr, TestPlan, StationId, Execution Time, and many more). However, we strongly recommend that you add additional fields to the test sequence to allow for more targeted filtering. The Report Analyzer can also group by self-assigned properties! (see examples "Nominal Current", "Nominal Frequency", "Number of Poles", etc.).
Even if no structuring of the data set is desired, at least one group must still be created so that the program can perform the analyses. In this case, simply click Add Group(s) without filtering to create a default group.
The figure below shows the layout of the Grouping Tab. This consists of three sections:
Upper half: Overview of all available Executions, either as a flat list or in groups, depending on how they are arranged. Here you can select the executions that should be created as new group(s).
Bottom left: Overview of all already created groups (Active Groups). These can be moved, renamed or deleted. A subsequent addition of executions is not possible.
Bottom right: Overview of all executions of the group selected on the left. If necessary, executions can be deleted here. Results will be recalculated.
Procedure
The upper view is configured using grouping (drag a column header in the grouping area above the table header) or filtering as desired
via the Add Group(s) button all groups or executions selected via checkbox will be added as new group(s). If nothing is selected, all currently visible executions and groups will be added.
In the list of active groups, the group name can be adjusted (F2) and the order can be changed. In the order defined here, the analyses will later be executed, displayed and structured in the export. The order can also be changed later. The selected groups are removed via the "Trash" button after confirmation.
A (re)calculation of the analysis is started automatically when leaving the grouping page.
Simplified group naming can be activated in the application settings. In this case, only the values without the column names are used as groups. (e.g. instead of "ProductId: xyz" only "xyz").
Groups can be added or deleted later. The analysis and graphs will then be recalculated. If the corresponding group is no longer available, it will be removed from the graph.
Individual executions can be selectively deleted from already created groups. When leaving the grouping page, a new analysis will be started.
Not all executions have to be assigned to a group.
Only executions in groups are analyzed.
An execution can also be assigned to several groups and will then be considered in both groups during the analysis.
The button Assignments can be used to display which executions are not assigned to any or several groups (see Manage groups).
Attention When the program is closed, all filter settings are saved and automatically applied when the program is restarted. If you have set a filter when creating the group, when loading the next dataset it may happen that all executions over the filter are hidden. If you remove the filter, the data will be visible again.
Here, the groups are formed individually via the filter settings. The filter is set as required (see chapter Filtering) and the visible executions are added as a group. The process is repeated until all desired groups are present.
This procedure is recommended for the following scenarios:
No grouping necessary (corresponds to the creation of exactly one group).
Grouping by time ranges
Complex groupings that cannot be realized via automated grouping
Procedure
the executions are filtered so that only the required entries are visible Only these entries are included in the group
click on Add Group(s) to create a group with a default name (e.g. Group_0)
with F2 the group can be renamed
steps 1-3 can be repeated as often as you like to create multiple groups
The fastest way to structure the data set is to use automated grouping. To do this, simply drag the column header to be grouped into the grouping area above the table header.
Procedure
Drag and drop the desired column header into the empty area above the table ("Drag a column header here to group by that column.")
The program now groups the data according to the values of the column you selected. The table immediately displays all the groups that have been formed as a preview. Use the arrow symbol to view the contents of the group.
Click Add Group and the groups are created with a default name and added to the list under Active Groups. The group names can be renamed with F2 if necessary.
Tip: A combined grouping is also possible. Simply drag another column header into the grouping area and the second level of the grouping is immediately visible in the preview:
Automatic groupings can also be filtered, allowing you to combine both approaches. For example, you can group by ProductID and then use filters to select only some of the automatically created groups:
The button Assignments allows an analysis of the distribution of executions to groups.
A table lists all executions and informs about the number of assignments and the assigned groups to an execution. This makes it easy to detect wrong or unwanted multiple assignments or even forgotten groups.
For this purpose the table offers additional predefined filters:
All: Shows all imported executions
Duplicates: Displays only those executions that are assigned to multiple groups
Unrelated: Displays only those executions that are not assigned to a group
In the main view of the Grouping Tab there is also the possibility to remove unassigned executions and thus free memory. To do this, just click on Cleanup.
Creating groups can be very time-consuming, especially creating complex filter expressions using the Filter Editor. Therefore, it is possible to save the current grouping visible in the Executions as a filter set and to load and apply this set again later.
Apply filter set If a filter set is to be applied, it only needs to be selected from the drop-down list. The settings take effect immediately. Existing other filters are reset.
Manage Filter Sets With the button Manage Filter-Sets the filter sets can be managed:
Export: The selected filter set is exported as an XML file and can be copied to other computers, for example Import: A previously exported XML filter set can be imported Delete: The selected filter set is removed from the list
Save filter set To save, simply click on the Save Filter Set button and enter the name of the set. The set will automatically be saved in the app data and displayed in the drop down list.
