Overview
QA Bin Definitions
Parameters :terminology (e.g. NCIMTH) and :version (e.g. latest) are allowed in the queries.\
TODO:
- Look up definitions of bins (e.g. logically what do they do, we'll need descriptions too
- Write up description
- Write and test query (with :terminology/:version interpolations)
REQUIRED bins
nci_merge: Merged SCUI current version NCI atoms
select distinct a.id conceptId from concepts a, concepts_atoms b, atoms c where a.terminology = :terminology and a.id = b.concepts_id and b.atoms_id = c.id and c.terminology='NCI' group by a.id having count(distinct c.conceptId)>1
nci_sub_split: Split SCUI current version NCI (or sub-source) atoms
TBD: select distinct a.id conceptId from concepts a, concepts_atoms b, atoms c where a.terminology = :terminology and a.id = b.concepts_id and b.atoms_id = c.id and c.terminology in (select terminology from root_terminologies where family = 'NCI' and terminology != 'NCI') group by a.id having count(distinct c.conceptId)>1
sct_sepfnpt: SNOMED concept clusters where the FN and PT terms are separated
-- ATOM Terminology = SNOMEDCT_US? Because only project-terminology concepts are visible through UI... SELECT DISTINCT c.id conceptId1, c1.id conceptId2 FROM concepts c, concepts_atoms ca, atoms a, concepts c1, concepts_atoms ca1, atoms a1 WHERE c.terminology = :terminology AND c1.terminology = :terminology AND a.terminology = 'SNOMEDCT_US' AND a1.terminology = 'SNOMEDCT_US' AND c.id = ca.concepts_id AND ca.atoms_id = a.id AND c1.id = ca1.concepts_id AND ca1.atoms_id = a1.id AND a.termType = 'FN' AND a1.termType = 'PT' AND a.conceptId = a1.conceptId AND c.id != c1.id;
cdsty_coc: Find concepts with Clinical Drug STY and any other STY
SELECT DISTINCT c.id conceptId FROM concepts c, concepts_semantic_type_components cs WHERE c.terminology = :terminology AND c.id = cs.concepts_id AND cs.concepts_id IN (SELECT cs.concepts_id FROM concepts c, concepts_semantic_type_components cs, semantic_type_components s WHERE c.terminology = :terminology AND c.id = cs.concepts_id AND cs.semanticTypes_id = s.id AND s.semanticType = 'Clinical Drug') GROUP BY cs.concepts_id HAVING COUNT(cs.concepts_id) > 1
multsty: concepts with more than 3 STYs|
SELECT DISTINCT cs.concepts_id conceptId FROM concepts c, concepts_semantic_type_components cs, semantic_type_components s WHERE c.terminology = :terminology AND c.id = cs.concepts_id AND cs.semanticTypes_id = s.id GROUP BY c.id HAVING COUNT(c.id) > 3
styisa: One STY is an ancestor of another in the STY isa hierarchy
SemanticType relationships?
sfo_lfo: Short form in one concept, long form in another
ShortForm/LongForm of what object?
deleted_cui: CUIs that are going away - will need bequeathal rel
unpublishable NCIMTH concepts?
NON-required bins
- rxnorm_merge: RXCUI Merges
Same as nci_merge but with RXNORM
select a.id clusterId, a.id conceptId from concepts a, concepts_atoms b, atoms c where a.terminology = :terminology and a.id = b.concepts_id and b.atoms_id = c.id and c.terminology='RXNORM' group by a.id having count(distinct c.conceptId)>1
- cbo_merge: Merged SCUI current version CBO atoms
Same as nci_merge but with CBO
select a.id clusterId, a.id conceptId from concepts a, concepts_atoms b, atoms c where a.terminology = :terminology and a.id = b.concepts_id and b.atoms_id = c.id and c.terminology='CBO' group by a.id having count(distinct c.conceptId)>1
mdr_merge: Merged SDUI current version MDR atoms
SELECT c.id clusterId, c.id conceptId FROM concepts c, concepts_atoms ca, atoms a WHERE c.terminology = :terminology AND c.id = ca.concepts_id AND ca.atoms_id = a.id AND a.terminology = 'MDR' GROUP BY c.id HAVING COUNT(DISTINCT a.descriptorId) > 1
pdq_merge: Merged SDUI current version PDQ atoms
SELECT c.id clusterId, c.id conceptId FROM concepts c, concepts_atoms ca, atoms a WHERE c.terminology = :terminology AND c.id = ca.concepts_id AND ca.atoms_id = a.id AND a.terminology = 'PDQ' GROUP BY c.id HAVING COUNT(DISTINCT a.descriptorId) > 1
- sct_sepfnpt: SNOMED concept clusters where the FN and PT terms are separated
- DUPLICATE of above
rxnorm_split: RXCUI splits
nci_pdq_merge: Concepts containing current version NCI and current version PDQ atoms
SELECT DISTINCT c.id conceptId FROM atoms a, atoms a1, concepts c, concepts c1, concepts_atoms ca, concepts_atoms ca1, terminologies t, terminologies t1 WHERE c.terminology = :terminology AND c1.terminology = :terminology AND c.id = ca.concepts_id AND ca.atoms_id = a.id AND c1.id = ca1.concepts_id AND ca1.atoms_id = a1.id AND a.terminology = 'NCI' AND a.terminology = t.terminology AND a.version = t.version AND t.current = TRUE AND a1.terminology = 'PDQ' AND a1.terminology = t1.terminology AND a1.version = t1.version AND t1.current = TRUE AND c.id = c1.id
nci_sct_merge: Concepts containing current version NCI and current version SNOMEDCT atoms
SELECT DISTINCT c.id conceptId FROM atoms a, atoms a1, concepts c, concepts c1, concepts_atoms ca, concepts_atoms ca1, terminologies t, terminologies t1 WHERE c.terminology = :terminology AND c1.terminology = :terminology AND c.id = ca.concepts_id AND ca.atoms_id = a.id AND c1.id = ca1.concepts_id AND ca1.atoms_id = a1.id AND a.terminology = 'NCI' AND a.terminology = t.terminology AND a.version = t.version AND t.current = TRUE AND a1.terminology = 'SNOMEDCT_US' AND a1.terminology = t1.terminology AND a1.version = t1.version AND t1.current = TRUE AND c.id = c1.id
ambig_no_pn: Ambiguous concepts where at least one has no MTH/PN nor NCIMTH/PN (sepstring)
--Ambiguous concept = concepts that share atoms that have the same case-insensitive name -- This is used by several of the below queries, so it should be created as a VIEW CREATE VIEW ambig_concepts AS SELECT DISTINCT c1.id conceptId1, c2.id conceptId2 FROM concepts c1, concepts_atoms ca1, atoms a1, concepts c2, concepts_atoms ca2, atoms a2 WHERE c1.terminology = 'NCIMTH' AND c2.terminology = 'NCIMTH' AND c1.id = ca1.concepts_id AND ca1.atoms_id = a1.id AND c2.id = ca2.concepts_id AND ca2.atoms_id = a2.id AND c1.id != c2.id AND a1.lowerNameHash = a2.lowerNameHash AND a1.id != a2.id AND a1.publishable = TRUE AND a2.publishable = TRUE
--Use Pre-programmed View ambig_concepts SELECT conceptId1, conceptId2 FROM ambig_concepts WHERE NOT ambig_concepts.conceptId1 IN (SELECT ca.concepts_id FROM concepts_atoms ca, atoms a WHERE ca.atoms_id = a.id AND a.terminology IN ('NCIMTH' , 'MTH') AND a.termType = 'PN') AND NOT ambig_concepts.conceptId2 IN (SELECT ca.concepts_id FROM concepts_atoms ca, atoms a WHERE ca.atoms_id = a.id AND a.terminology IN ('NCIMTH' , 'MTH') AND a.termType = 'PN')
ambig_no_rel: Ambiguous concepts that lack an approved REL
--Use Pre-programmed View ambig_concepts SELECT conceptId1, conceptId2 FROM ambig_concepts WHERE NOT (conceptId1 , conceptId2) IN (SELECT cr.from_id, cr.to_id FROM concept_Relationships cr WHERE cr.publishable = TRUE AND cr.workflowStatus in ('READY_FOR_PUBLICATION','PUBLISHED'))
pn_pn_ambig: Identical (same SUI) PN's in multiple concepts
multiple_pn: Concepts with multiple MTH/PN atoms
SELECT DISTINCT c.id conceptId FROM concepts c, atoms a, concepts_atoms ca WHERE c.terminology = :terminology AND c.id = ca.concepts_id AND ca.atoms_id = a.id AND a.terminology in ('MTH', 'NCIMTH') AND a.termType = 'PN' GROUP BY c.id HAVING COUNT(c.id) > 1
pn_no_ambig: Concept has MTH/PN atom but no ambiguous string
ambig_pn: MTH/PN atom is ambiguous but has no matching ambiguous string
pn_orphan: MTH/PNs on their own
--Concepts whose only publishable atoms are MTH/PN or NCIMTH/PN
nosty: No STY
SELECT DISTINCT c.id conceptId FROM concepts c LEFT JOIN concepts_semantic_type_components cs ON (c.id = cs.concepts_id) WHERE c.terminology= :terminology AND cs.semanticTypes_id IS NULL
missing_sty: Reviewed concepts without releasable Semantic Types
SELECT DISTINCT c.id conceptId FROM concepts c WHERE c.terminology = :terminology AND c.publishable = TRUE AND c.workflowStatus IN ('READY_FOR_PUBLICATION' , 'PUBLISHED') AND NOT c.id IN (SELECT c.id FROM concepts c, concepts_semantic_type_components cs, semantic_type_components s WHERE c.terminology = :terminology AND c.id = cs.concepts_id AND cs.semanticTypes_id = s.id AND s.publishable = TRUE)
cbo_chem: Current version CBO concepts with CHEM STYs
-- to find chemical semantic types (select distinct semanticTypeCategoryMap sty from projects a, ProjectJpa_semanticTypeCategoryMap b where a.id = b.ProjectJpa_id and semanticTypeCategoryMap_KEY = 'chemical' and a.terminology = :terminology); -- Rick update 3/2/2017 (select distinct semanticTypeCategoryMap_KEY sty from projects a, ProjectJpa_semanticTypeCategoryMap b where a.id = b.ProjectJpa_id and semanticTypeCategoryMap = 'chem' and a.terminology = :terminology)
go_chem: Current version GO concepts with CHEM STYs
SELECT DISTINCT c.id conceptId FROM concepts c, concepts_semantic_type_components cs, semantic_type_components s, terminologies t WHERE c.terminology = 'GO' AND c.terminology = t.terminology AND c.version = t.version AND t.current = TRUE AND c.id = cs.concepts_id AND cs.semanticTypes_id = s.id AND s.semanticType IN ((SELECT DISTINCT semanticTypeCategoryMap_KEY sty FROM projects a, ProjectJpa_semanticTypeCategoryMap b WHERE a.id = b.ProjectJpa_id AND semanticTypeCategoryMap = 'chem' AND a.terminology = :terminology))
mdr_chem: Current version MDR concepts with CHEM STYs
See above
true_orphan: Concepts with no releaseable relationships to any other concept
SELECT c.id conceptId FROM concepts c WHERE c.terminology = :terminology AND NOT c.id IN (SELECT c.id FROM concepts c, concept_relationships cr WHERE cr.from_id = c.id AND cr.publishable = TRUE)
deleted_cui_split: Complex split/merge case CUIs that are going away - will need bequeathal
- split_demotions: Demotions overlapping with a bad merge that was split
- nci_pt_mrg: Merged current version NCI/PT atoms
- mxsuppr: Concepts with identical LUI atoms that have mixed suppressibility
.~ Mixed status concepts 196 Sat Jan 7 09:19:47 2017
~ Suppressible preferred name (level 0,9) 0 Sat Jan 7 09:19:58 2017