2.7.3.4. Search, open, and delete saved clusterings

Purpose

In the Clustering Overview, you can view a completed clustering result, read its summary, open it for a detailed analysis, or permanently delete an outdated result.

Requirements

To access saved clustering results, the logged-in account must have the "clustering-view" permission.
For "Generate names" and "Delete," clustering-manage is required. Accounts with read-only access can still open a result.
The desired clustering must already be complete and listed in the "Clusterings" section.

Procedure

"Filter" narrows down the overview without distinguishing between uppercase and lowercase letters. A saved result can be found in three ways: by the clustering name, the saved catalog name, and the catalog's display name shown on the map. The corresponding clustering map [2] serves as an overview for operation and opens the saved result directly.

A filtered overview displays the selected saved result, including the source, configuration, and test bench.

Cluster Name

Use the saved clustering name [1] if you know exactly which result you are looking for. This is the most precise filter for distinguishing between individual runs of the same catalog.

The filter by clustering name narrows the results down to a specific run.

Catalog Name

Using the saved catalog name [1], you can group together the saved results from a technical source catalog. This is useful when you want to review multiple runs of the same source together.

The filter by catalog name groups together all saved results from a source catalog.

Catalog Display Name

Use the human-readable display name of the catalog [1] if you are familiar with this term. It returns the same saved results as the saved catalog name, without requiring you to know the technical catalog identifier.

The filter by catalog display name returns the same results without the technical catalog identifier.

In the clustering details view, you can analyze the results. The title [1] indicates the currently loaded result. Return to the overview if you want to select a different result or manage its lifecycle .

The title refers to the saved result, which is analyzed in the detail view.

The "More" [1] menu lists the actions available for the selected card. “Open” leaves the result unchanged and opens its detail view. “Delete” [2] initiates the deliberate, permanent removal of results that are outdated, duplicated, or no longer needed for other reasons.

The card menu separates normal access to the result from the lifecycle actions that are available only to administrators .

Options and Decisions in Operations

Saved Clusterings – Filter Options and Lifecycle Actions

Control element or State	Operating Instructions
"Filter"	Searches the cards of saved clusterings without distinguishing between uppercase and lowercase letters. The filter is optional and initially empty; delete it to view the complete overview again. The search options are described in the following lines .
Cluster Name	Finds the saved clustering name, either in full or in part. Use this to find a known run, especially when multiple results come from the same catalog.
Catalog Name	Finds all or part of the internal catalog name stored using clustering. Use this to summarize the stored results of a source catalog if you know its technical name.
Display name of the catalog	Finds the catalog name shown on the clustering maps, either in full or in part. Use this option if you know the catalog's display name, but not its stored name.
"Clusterings"	Contains saved results, no running jobs. The cards are sorted by creation date, with the most recent result first, so that current tasks appear at the top of the overview. A card remains here regardless of whether the job list is cleared, until an administrator deletes its data.
Clustering Map	Opens the specified result. The name of the card is the clustering name or, if none was specified, the display name of the catalog. Distinguish between similar runs based on their name and catalog.
Date Created and Source Catalog	Show when the result was generated and which catalog its parts come from. This information helps prevent you from accidentally opening or deleting a result with the same name from a different source.
"Preset," "Granularity," "Max depth" and enabled name-generation options	Summarize the submitted clustering configuration, provided it is saved along with the result . The granularity is displayed as a percentage ; the maximum depth is either a number of levels or “unlimited levels”; Name generation and directory name generation appear only if they were enabled. Use these values when deciding whether a result is suitable for review or needs to be replaced.
"Clusters," "Parts," "Done," "To check," and "Open"	Summarize the loaded results . The three inspection counters show how much rework is still needed on the results; together, they describe the visible distribution of part statuses.
Histogram	Shows the distribution of cluster sizes. This allows you to determine, before conducting a more detailed analysis, whether a result is characterized by very small, very large, or uneven clusters.
"More"	Opens the action menu for this saved clustering. It serves as the entry point for "Open" and, for administrators, for "Generate names" and "Delete."
"Open"	Opens the selected saved result without modifying the saved data. Available to users with the clustering-view.
"Generate names"	Opens a dialog to start a custom name-generation job for the selected result. The dialog can generate directory names for all tree nodes and is available only with clustering-manage. You can manage this job in the name-generation workflow.
"Delete"	Available only with clustering-manage. Requests confirmation to permanently delete the data for the selected result. Do not use this action until you are certain that the result is obsolete and neither the history nor the analysis is needed anymore.
Confirmation: "Delete clustering"	Specify the result in question and note that the deletion cannot be undone . “Cancel” leaves the saved result unchanged; “Delete” permanently removes it from the overview and the backend’s data storage.

Result

The administrator can narrow down and identify a saved clustering, work with the detailed results, and remove outdated data only after express confirmation by a manager. The overview remains the authoritative list of saved results.

Tips and Troubleshooting

If the filtered list is empty, the search text may not match the clustering name, the catalog name, or the display name of the catalog. Clear or refine the filter before assuming the result is unavailable.
If "Delete" is missing from the "More" menu, the logged-in account does not have the "clustering-manage" permission; however, you can still open the result using "clustering-view."
Permanent deletion cannot be undone in the clustering interface. Select “Cancel” while the confirmation dialog is still open if you need to review the result.

Prev	Up	Next
2.7.3.3. Monitor and manage clustering jobs	Home	2.7.3.5. Search and analyze clustering results