Whenever I want to pivot a table in Hive, I collect key:value
pairs to a map and then reference each key in the next level, creating new columns. This is the opposite of that.
Query:
select a.userid, y.new_id
from (
select new_id, fruit_name, fruit_code
from (
select new_id, map("apple_id", apple_id
, "mango_id", mango_id
, "grape_id", grape_id
, "peach_id", peach_id) as fruit_map
from table_2 ) x
lateral view explode(fruit_map) exptbl1 as fruit_name, fruit_code ) y
join table_A a
on (y.fruit_code=a.property_id)
Output:
0001 N456098
0031 N002345
0008 N129087
00013 N109876