nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Own type based criterion in advanced search cluster strategies

jmeabe

Hi!

I'm configuring the advanced search cluster strategies. If I create a cluster strategy based on a documentum custom attribute there is no problem, but if I do it using an own attribute no results are displayed in the cluster tree view: Not enough relevant data to compute clusters. And of course I have created some documents with this attribute setted. Seems like something lacks in my own attributes. Should I register them somewhere apart from the strategies definiton in clusterstrategies_modifications.xml file?

Thanks in advance!

Find more posts tagged with

Comments

annetteb

Hi,

The clusters are computed to group results and show the most significant metadata about the results. So, by default, a threshold value is set to indicate a minimum number of results to compute a meaningful cluster. The message "Not enough relevant data to compute clusters" just means that. When you say that you created "some" documents, maybe you don't have enough documents?

Note that clusters are not facets: they are not designed to provide an exhaustive view of the result set.

Regarding your attributes, what is the difference you make between "documentum custom attributes" and "my own attributes"? Clustering would not work on custom document properties like the Author property of a PDF document.

Regards,

Annette

ByDsr

Hi,

I have the same problem like jmeabe.

I'm configuring the advanced search cluster strategies, I have a documental type, if I configure an attribute of a supertype, and that supertype is an original documentum type, no problem, but y try to show attributes of my custom documental type.

I put the threshold =1, and nothing happens, webtop show me the same message:

"Not enough relevant data to compute clusters"

Have an idea?

Thanks...

DCTM_Guru

As Annetteb has suggested, how many sample documents do you have in the system that server can use to generate a cluster?

ByDsr

About 8

DCTM_Guru

I would say you need about 25+ with enough variations in the values, such that server can identify a meaningful cluster. All documents have same value or all documents having different (distinct) values is not going to generate a good cluster. You need a good sampling that has some repetitiveness as well as distinct values for it to work.

ByDsr

Ok, thanks Johnny Gee, I'm going to try it, because I have all documents with the same values... I think

annetteb

Hi,

I don't know what is exactly your requirements but maybe the facets would provide what you're looking for.

If you're interested, have a look at this document: https://community.emc.com/docs/DOC-12983.

Best regards,

Annette

lastnitescurry

@ ByDsr, jmeabe

Where are you executing your searches from? Quick Search or Advanced Search... I assume we are talking about Webtop!!!

The type which defines the cluster attribute must be in the search DQL FROM clause. Do not be fooled by the attribute being populated in Webtop Search Results columns list... that is populated after the search has been executed.

For example: If you have a cluster strategy for attribute bdd_reviewer in type sop_doc, then you must search on type

sop_doc to get the cluster for bdd_reviewer to populate. i.e. you must use Advanced Search rather than Quick Search

ByDsr

Thanks Brian Dinneen, I'm executing my searches from Advanced Search of Webtop, and the problem is the results don´t show in the clusters... this cluster strategies are attributes of the documental type.

lastnitescurry

confused!!! are you saying

thanks my problem is solved
or thanks but I still do not know what to do?

MuhammadAli

I'm not sure it would help but one thing I have noticed in clustering is that, when your search results contain lets say 50 sop_doc documents and clustering is done on say sop_type attribute. Then cluster would only be created not only when the threshold is met but also when results have at least two distinct sop_type attribute values. If all your 50 documents have same sop_type value, then the cluster would not show.