przemyslaw.jacewicz
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

update sql query to return distinct pids

This commit is contained in:
Claudio Atzori 2023-06-12 09:47:45 +02:00
parent da7b66c542
commit e1409ffe80
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/sql/queryOpenOrgsForProvision.sql
View File

@ -15,7 +15,10 @@ SELECT
'OpenOrgs Database' AS collectedfromname, 'OpenOrgs Database' AS collectedfromname,
o.country || '@@@dnet:countries' AS country, o.country || '@@@dnet:countries' AS country,
'sysimport:crosswalk:entityregistry@@@dnet:provenance_actions' AS provenanceaction, 'sysimport:crosswalk:entityregistry@@@dnet:provenance_actions' AS provenanceaction,
array_remove(array_cat(array_agg(DISTINCT i.otherid || '###' || i.type || '@@@dnet:pid_types'), array_agg(DISTINCT idup.otherid || '###' || idup.type || '@@@dnet:pid_types')), NULL) AS pid, ARRAY(SELECT DISTINCT pid FROM unnest(array_cat(
array_agg(DISTINCT i.otherid || '###' || i.type || '@@@dnet:pid_types'),
array_agg(DISTINCT idup.otherid || '###' || idup.type || '@@@dnet:pid_types')
)) as t(pid) where pid IS NOT NULL),
(array_remove(array_cat(ARRAY[o.ec_legalbody], array_agg(od.ec_legalbody)), NULL))[1] AS eclegalbody, (array_remove(array_cat(ARRAY[o.ec_legalbody], array_agg(od.ec_legalbody)), NULL))[1] AS eclegalbody,
(array_remove(array_cat(ARRAY[o.ec_legalperson], array_agg(od.ec_legalperson)), NULL))[1] AS eclegalperson, (array_remove(array_cat(ARRAY[o.ec_legalperson], array_agg(od.ec_legalperson)), NULL))[1] AS eclegalperson,
(array_remove(array_cat(ARRAY[o.ec_nonprofit], array_agg(od.ec_nonprofit)), NULL))[1] AS ecnonprofit, (array_remove(array_cat(ARRAY[o.ec_nonprofit], array_agg(od.ec_nonprofit)), NULL))[1] AS ecnonprofit,