Investigate star:E13_sdhss_P17 star

[lu-pl]

The construct query for generating star:E13_sdhss_P17 starlegs assertions finds the common connector in the star pattern and isolates the connected E13 nodes that do not have P14/P17 assertions pointing to the same objects as the initial E13. The problem with star:E13_sdhss_P17 is that lots of the E13 nodes turn out to actually have P14/P17 assertions already attached to them.

construct {
    ?o crm:P14_carried_out_by ?agent .
    ?o crm:P17_was_motivated_by ?source .
}
where {
    ?e13_initial a star:E13_sdhss_P17 ;
    	crm:P140_assigned_attribute_to ?common .

    optional {?e13_initial crm:P14_carried_out_by ?agent}
    optional {?e13_initial crm:P17_was_motivated_by ?source}

    ?common ^crm:P140_assigned_attribute_to ?o .
    filter (?o != ?e13_initial)

    minus {?o crm:P14_carried_out_by ?agent}
    minus {?o crm:P17_was_motivated_by ?source}
}

The reason why the connected E13s do not get subtracted in the MINUS clauses is that their P14/P17 assertions point to other objects than the initial E13. So running the above construct query would generate additional P14/P17 for almost all matches.

This should be investigated.