D-Net
/
dnet-hadoop
13
4
Fork 10
Code Issues 14 Pull Requests 8 Projects 2 Releases 1 Wiki Activity

added eoscifguidelines in the result view; removed compute statistics statements

This commit is contained in:
Claudio Atzori 2023-04-11 10:43:32 +02:00
parent 932d07d2dd
commit a2dcb06daf
1 changed files with 4 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/hive/oozie_app/lib/scripts/postprocessing.sql
View File

@ -1,19 +1,10 @@
DROP VIEW IF EXISTS ${hiveDbName}.result;
CREATE VIEW IF NOT EXISTS ${hiveDbName}.result as
select id, originalid, dateofcollection, title, publisher, bestaccessright, datainfo, collectedfrom, pid, author, resulttype, language, country, subject, description, dateofacceptance, relevantdate, embargoenddate, resourcetype, context, externalreference, instance, measures, processingchargeamount from ${hiveDbName}.publication p
select id, originalid, dateofcollection, title, publisher, bestaccessright, datainfo, collectedfrom, pid, author, resulttype, language, country, subject, description, dateofacceptance, relevantdate, embargoenddate, resourcetype, context, externalreference, instance, measures, processingchargeamount, eoscifguidelines from ${hiveDbName}.publication p
union all
select id, originalid, dateofcollection, title, publisher, bestaccessright, datainfo, collectedfrom, pid, author, resulttype, language, country, subject, description, dateofacceptance, relevantdate, embargoenddate, resourcetype, context, externalreference, instance, measures, processingchargeamount from ${hiveDbName}.dataset d
select id, originalid, dateofcollection, title, publisher, bestaccessright, datainfo, collectedfrom, pid, author, resulttype, language, country, subject, description, dateofacceptance, relevantdate, embargoenddate, resourcetype, context, externalreference, instance, measures, processingchargeamount, eoscifguidelines from ${hiveDbName}.dataset d
union all
select id, originalid, dateofcollection, title, publisher, bestaccessright, datainfo, collectedfrom, pid, author, resulttype, language, country, subject, description, dateofacceptance, relevantdate, embargoenddate, resourcetype, context, externalreference, instance, measures, processingchargeamount from ${hiveDbName}.software s
select id, originalid, dateofcollection, title, publisher, bestaccessright, datainfo, collectedfrom, pid, author, resulttype, language, country, subject, description, dateofacceptance, relevantdate, embargoenddate, resourcetype, context, externalreference, instance, measures, processingchargeamount, eoscifguidelines from ${hiveDbName}.software s
union all
select id, originalid, dateofcollection, title, publisher, bestaccessright, datainfo, collectedfrom, pid, author, resulttype, language, country, subject, description, dateofacceptance, relevantdate, embargoenddate, resourcetype, context, externalreference, instance, measures, processingchargeamount from ${hiveDbName}.otherresearchproduct o;
ANALYZE TABLE ${hiveDbName}.datasource COMPUTE STATISTICS;
ANALYZE TABLE ${hiveDbName}.organization COMPUTE STATISTICS;
ANALYZE TABLE ${hiveDbName}.project COMPUTE STATISTICS;
ANALYZE TABLE ${hiveDbName}.publication COMPUTE STATISTICS;
ANALYZE TABLE ${hiveDbName}.dataset COMPUTE STATISTICS;
ANALYZE TABLE ${hiveDbName}.otherresearchproduct COMPUTE STATISTICS;
ANALYZE TABLE ${hiveDbName}.software COMPUTE STATISTICS;
ANALYZE TABLE ${hiveDbName}.relation COMPUTE STATISTICS;
select id, originalid, dateofcollection, title, publisher, bestaccessright, datainfo, collectedfrom, pid, author, resulttype, language, country, subject, description, dateofacceptance, relevantdate, embargoenddate, resourcetype, context, externalreference, instance, measures, processingchargeamount, eoscifguidelines from ${hiveDbName}.otherresearchproduct o;