Hi:
Recently upgraded hive version to 1.1.0 from 0.10.0 The same script get
different results in two environments.By step by step,Problem locating at
LATERAL VIEW explode(split(yhd, ','))
Business script:
select count(*) FROM default.track_exps m LATERAL VIEW
explode(split(m.attachedinfo, ',')) mytable1 AS productid WHERE ds =
'2015-10-14';
hive0.10 result *1006387352*
hive1.1.0 result *1008900109*,
Difference quantity *2512757*
By looking at the data, it is found that there is' 1,2 ',' 1,2, 'type of
data,Guess the problem come from ’,’,run select count(*) from
default.track_exps WHERE ds = '2015-10-14' and
substring(attachedinfo,length(attachedinfo))=','; result 2512757,the same
to difference quantity。
Re testing:select * from default.dual LATERAL VIEW explode(split('1,',
',')) mytable1 AS productid;
result:
Hive0.10 *test 1*
(1 row)
----------------------------------------------------------------------------------------------------------------------------------------------------------
Hive1.1.0 *test 1*
*test*
(2 rows)
Test results show that the two versions of the '1,2,' type of data
processing methods are not the same,Please see if it's HIVE's Bug.
Thanks!