如何使用存储过程在 sql server 中 PIVOT 表？

Question

My Stg table looks like,
Com_key  metric_name  metric_value  metric_date
   A        bugs           1        2020-06-09T16:26:41+0000
   A        bugs           1        2020-06-09T16:31:55+0000
   B        bugs           0        2020-06-18T13:59:28+0000
   B        bugs           1        2020-06-18T14:11:13+0000
   C        bugs           1        2020-06-15T14:27:32+0000
My fact table should look like,

com_key    bugs_cnt   metric_date
   A           1      2020-06-09T16:31:55+0000  
   B           0      2020-06-18T13:59:28+0000
   B           1      2020-06-18T14:11:13+0000
   C           1      2020-06-15T14:27:32+0000

I need to write a stored procedure to PIVOT table in sql server for the below explained scenario. When duplicate record comes on same day I need to consider only the last modified date. In the above example there is a duplicate record for com_key A with same metric_date and different time and I need to consider only the latest updated value, but at the same time need to retain the com_keys with diff metric values which comes on the same day with different time.

Thanks in advance.

Answer 1

What pivot? this looks like filtering:

select Com_key, metric_name, metric_value, max(metric_date) as metric_date
from (select t.*,
             row_number() over (partition by Com_key, metric_name, metric_value, convert(date, metric_date) order by metric_date desc) as seqnum
      from t
     ) t
where seqnum = 1;

I also strongly discourage from putting this into a stored procedure. If you want to store this as code, a view is quite sufficient.