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

added stats computation on the graph hive DB tables

This commit is contained in:
Claudio Atzori 2021-12-02 10:44:10 +01:00
parent cfa4560769
commit 3b19821f3c
1 changed files with 9 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
dhp-workflows/dhp-graph-mapper/src/main/resources/eu/dnetlib/dhp/oa/graph/hive/oozie_app/lib/scripts/postprocessing.sql
View File

@ -8,3 +8,12 @@ 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 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 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;