forked from D-Net/dnet-hadoop
using distinct apcs per publication to avoid huge sums
This commit is contained in:
parent
2e4cab026c
commit
316d585c8a
|
@ -56,7 +56,7 @@ join ${stats_db_name}.result res on res.id=r.id;
|
|||
DROP TABLE IF EXISTS ${stats_db_name}.result_apc purge;
|
||||
|
||||
create table if not exists ${stats_db_name}.result_apc STORED AS PARQUET as
|
||||
select r.id, r.amount, r.currency
|
||||
select distinct r.id, r.amount, r.currency
|
||||
from (
|
||||
select substr(r.id, 4) as id, cast(inst.processingchargeamount.value as float) as amount, inst.processingchargecurrency.value as currency
|
||||
from ${openaire_db_name}.result r lateral view explode(r.instance) instances as inst) r
|
||||
|
|
Loading…
Reference in New Issue