Aggregation, mathematical function & GROUP BY in a single query in Hive - hadoop

Question

I have Table T1 with the below schema:
job_id job_name queue memory cores start_time end_time
job_1234 ABC A_user 51200 20 22-02-2018 22-02-2018
job_2345 ABC A_user 71680 30 22-02-2018 23-02-2018
I want the output to be:
ID f_queue f_job_name f_memory f_cores f_start_time f_end_time process_month
1 A_user ABC 120 50 22-02-2018 23-02-2018 201702
Where memory= (51200+71680/1024), cores=(20+30), ID and process_month are the static variables that I am passing to the hive script.
Is the below query the right one:
select
${ID},
job_id as f_queue,
job_name as f_job_name,
sum(memory)/1024 as f_memory,
sum(f_cores) as f_cores,
min(start_time) as f_start_time,
max(end_time) as f_end_time,
${process_month} as process_month
from T1 group by f_job_name,f_queue;

Answer 1

in group by you need to refer to actual column names i.e job_id,job_name not the alias names(i.e f_queue,f_job_name).
in your query you are doing group by job_id as f_queue i.e job_id field and job_name, when hive did group by on both job_id,job_name above records goes to different groups(job_1234,job_2345)
You need to change your query to
select
queue as f_queue,
job_name as f_job_name,
sum(memory)/1024 as f_memory,
sum(cores) as f_cores,
min(start_time) as f_start_time,
max(end_time) as f_end_time
from queue group by job_name,queue;
in the above query i'm doing group by on queue and job_name which will result
+----------+-------------+-----------+----------+---------------+-------------+--+
| f_queue | f_job_name | f_memory | f_cores | f_start_time | f_end_time |
+----------+-------------+-----------+----------+---------------+-------------+--+
| A_user | ABC | 120.0 | 50.0 | 22-02-2018 | 23-02-2018 |
+----------+-------------+-----------+----------+---------------+-------------+--+
Your query results:-
including job_name,job_id in group by clause
select
job_id as f_queue,
job_name as f_job_name,
sum(memory)/1024 as f_memory,
sum(cores) as f_cores,
min(start_time) as f_start_time,
max(end_time) as f_end_time
from queue group by job_name,job_id;
Output:-
+-----------+-------------+-----------+----------+---------------+-------------+--+
| f_queue | f_job_name | f_memory | f_cores | f_start_time | f_end_time |
+-----------+-------------+-----------+----------+---------------+-------------+--+
| job_1234 | ABC | 50.0 | 20.0 | 22-02-2018 | 22-02-2018 |
| job_2345 | ABC | 70.0 | 30.0 | 22-02-2018 | 23-02-2018 |
+-----------+-------------+-----------+----------+---------------+-------------+--+
Probably Your Required query Would be something like
select
${ID},
queue as f_queue,
job_name as f_job_name,
sum(memory)/1024 as f_memory,
sum(cores) as f_cores,
min(start_time) as f_start_time,
max(end_time) as f_end_time
${process_month} as process_month
from T1 group by queue,job_name;