lsmyrnaios
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

Merge pull request '[stats wf] Bug fix on indicators step' (#304) from antonis.lempesis/dnet-hadoop:beta into beta 

Reviewed-on: D-Net/dnet-hadoop#304
This commit is contained in:
Claudio Atzori 2023-06-07 16:49:09 +02:00
parent 01c67e697d fa24e2e18f
commit afb76ebf0f
1 changed files with 67 additions and 67 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

134
dhp-workflows/dhp-stats-update/src/main/resources/eu/dnetlib/dhp/oa/graph/stats/oozie_app/scripts/step16-createIndicatorsTables.sql
View File

@ -223,73 +223,6 @@ select id, count(id) as number_of_copies from result_instance group by id;
ANALYZE TABLE indi_result_no_of_copies COMPUTE STATISTICS; ANALYZE TABLE indi_result_no_of_copies COMPUTE STATISTICS;
---- Sprint 6 ---- ---- Sprint 6 ----
create table if not exists indi_pub_hybrid_oa_with_cc stored as parquet as
WITH hybrid_oa AS (
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_print as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_print != ""
UNION ALL
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_online as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_online != "" and (journal_is_in_doaj = FALSE OR journal_is_oa = FALSE)),
issn AS (
SELECT *
FROM (
SELECT id, issn_printed as issn
FROM datasource
WHERE issn_printed IS NOT NULL
UNION ALL
SELECT id,issn_online as issn
FROM datasource
WHERE issn_online IS NOT NULL ) as issn
WHERE LENGTH(issn) > 7)
SELECT DISTINCT pd.id, coalesce(is_hybrid_oa, 0) as is_hybrid_oa
FROM publication_datasources pd
LEFT OUTER JOIN (
SELECT pd.id, 1 as is_hybrid_oa from publication_datasources pd
JOIN datasource d on d.id=pd.datasource
JOIN issn on issn.id=pd.datasource
JOIN hybrid_oa ON issn.issn = hybrid_oa.issn
JOIN indi_result_has_cc_licence cc on pd.id=cc.id
JOIN indi_pub_gold_oa ga on pd.id=ga.id
where cc.has_cc_license=1 and ga.is_gold=0) tmp on pd.id=tmp.id;
ANALYZE TABLE indi_pub_hybrid_oa_with_cc COMPUTE STATISTICS;
create table if not exists indi_pub_bronze_oa stored as parquet as
WITH hybrid_oa AS (
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_print as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_print != ""
UNION ALL
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_online as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_online != "" and (journal_is_in_doaj = FALSE OR journal_is_oa = FALSE)),
issn AS (
SELECT *
FROM (
SELECT id, issn_printed as issn
FROM datasource
WHERE issn_printed IS NOT NULL
UNION ALL
SELECT id,issn_online as issn
FROM datasource
WHERE issn_online IS NOT NULL ) as issn
WHERE LENGTH(issn) > 7)
SELECT DISTINCT pd.id, coalesce(is_bronze_oa, 0) as is_hybrid_oa
FROM publication_datasources pd
LEFT OUTER JOIN (
SELECT pd.id, 1 as is_bronze_oa from publication_datasources pd
JOIN datasource d on d.id=pd.datasource
JOIN issn on issn.id=pd.datasource
JOIN hybrid_oa ON issn.issn = hybrid_oa.issn
JOIN indi_result_has_cc_licence cc on pd.id=cc.id
JOIN indi_pub_gold_oa ga on pd.id=ga.id
JOIN indi_pub_hybrid_oa_with_cc hy on hy.id=pd.id
where cc.has_cc_license=0 and ga.is_gold=0 and hy.is_hybrid_oa=0) tmp on pd.id=tmp.id;
ANALYZE TABLE indi_pub_bronze_oa COMPUTE STATISTICS;
create table if not exists indi_pub_downloads stored as parquet as create table if not exists indi_pub_downloads stored as parquet as
SELECT result_id, sum(downloads) no_downloads from openaire_prod_usage_stats.usage_stats SELECT result_id, sum(downloads) no_downloads from openaire_prod_usage_stats.usage_stats
join publication on result_id=id join publication on result_id=id
@ -376,6 +309,73 @@ FROM
ANALYZE TABLE indi_pub_gold_oa COMPUTE STATISTICS; ANALYZE TABLE indi_pub_gold_oa COMPUTE STATISTICS;
create table if not exists indi_pub_hybrid_oa_with_cc stored as parquet as
WITH hybrid_oa AS (
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_print as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_print != ""
UNION ALL
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_online as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_online != "" and (journal_is_in_doaj = FALSE OR journal_is_oa = FALSE)),
issn AS (
SELECT *
FROM (
SELECT id, issn_printed as issn
FROM datasource
WHERE issn_printed IS NOT NULL
UNION ALL
SELECT id,issn_online as issn
FROM datasource
WHERE issn_online IS NOT NULL ) as issn
WHERE LENGTH(issn) > 7)
SELECT DISTINCT pd.id, coalesce(is_hybrid_oa, 0) as is_hybrid_oa
FROM publication_datasources pd
LEFT OUTER JOIN (
SELECT pd.id, 1 as is_hybrid_oa from publication_datasources pd
JOIN datasource d on d.id=pd.datasource
JOIN issn on issn.id=pd.datasource
JOIN hybrid_oa ON issn.issn = hybrid_oa.issn
JOIN indi_result_has_cc_licence cc on pd.id=cc.id
JOIN indi_pub_gold_oa ga on pd.id=ga.id
where cc.has_cc_license=1 and ga.is_gold=0) tmp on pd.id=tmp.id;
ANALYZE TABLE indi_pub_hybrid_oa_with_cc COMPUTE STATISTICS;
create table if not exists indi_pub_bronze_oa stored as parquet as
WITH hybrid_oa AS (
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_print as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_print != ""
UNION ALL
SELECT issn_l, journal_is_in_doaj, journal_is_oa, issn_online as issn
FROM STATS_EXT.plan_s_jn
WHERE issn_online != "" and (journal_is_in_doaj = FALSE OR journal_is_oa = FALSE)),
issn AS (
SELECT *
FROM (
SELECT id, issn_printed as issn
FROM datasource
WHERE issn_printed IS NOT NULL
UNION ALL
SELECT id,issn_online as issn
FROM datasource
WHERE issn_online IS NOT NULL ) as issn
WHERE LENGTH(issn) > 7)
SELECT DISTINCT pd.id, coalesce(is_bronze_oa, 0) as is_hybrid_oa
FROM publication_datasources pd
LEFT OUTER JOIN (
SELECT pd.id, 1 as is_bronze_oa from publication_datasources pd
JOIN datasource d on d.id=pd.datasource
JOIN issn on issn.id=pd.datasource
JOIN hybrid_oa ON issn.issn = hybrid_oa.issn
JOIN indi_result_has_cc_licence cc on pd.id=cc.id
JOIN indi_pub_gold_oa ga on pd.id=ga.id
JOIN indi_pub_hybrid_oa_with_cc hy on hy.id=pd.id
where cc.has_cc_license=0 and ga.is_gold=0 and hy.is_hybrid_oa=0) tmp on pd.id=tmp.id;
ANALYZE TABLE indi_pub_bronze_oa COMPUTE STATISTICS;
create table if not exists indi_pub_hybrid stored as parquet as create table if not exists indi_pub_hybrid stored as parquet as
WITH gold_oa AS ( SELECT WITH gold_oa AS ( SELECT
issn_l, issn_l,