Find the first key by date field using sql and output also have other fields

偶尔善良 提交于 2019-12-24 19:24:14

问题


I want to query the first occurrence of every name according to the earliest date. The output should have the complete row. Please help me to write the query in sql.

Input:

  Name |    ID |    payment_date |  Pack
 ------+-------+-----------------+-------
    A  |    11 |    31-Jan       |  P
    C  |    13 |    31-Jan       |  Q
    B  |    2  |    31-Jan       |  R
    C  |    3  |    28-Jan       |  P
    D  |    23 |    29-Jan       |  Q
    B  |    11 |    30-Jan       |  R
    A  |    17 |    25-Jan       |  P
    C  |    13 |    26-Jan       |  Q
    D  |    17 |    2-Feb        |  R
    B  |    23 |    3-Feb        |  P
    A  |    45 |    4-Feb        |  Q
    B  |    3  |    5-Feb        |  R

Output:

  Name |    ID | payment_date | Pack
  -----+-------+--------------+-----
    A  |    17 | 25-Jan       | P
    B  |    11 | 30-Jan       | R
    C  |    13 | 26-Jan       | Q
    D  |    23 | 29-Jan       | Q

回答1:


You can use the min function, also assuming payment_date is a date type:

select Name, ID, min(payment_date), Pack from mytable
group by payment_date,Name, ID, Pack
order by Name

The downfall about this method is putting all of the fields in the group by.




回答2:


If your payment_date is a date data type, you can use not exists() like so:

select *
from t
where not exists (
  select 1
  from t i
  where i.Name = t.Name
    and i.payment_date < t.payment_date
    )

rextester demo (sql server): http://rextester.com/OKB46268

returns

+------+----+-------------+------+
| Name | Id | PaymentDate | Pack |
+------+----+-------------+------+
| A    | 17 | 2017-01-25  | P    |
| B    | 11 | 2017-01-30  | R    |
| C    | 13 | 2017-01-26  | Q    |
| D    | 23 | 2017-01-29  | Q    |
+------+----+-------------+------+



回答3:


You can also use Vertica's enhanced LIMIT clause:

WITH
-- input, don't use in real query
input(Name,ID,payment_date,Pack) AS (
          SELECT 'A',11,DATE '31-Jan-2017','P'
UNION ALL SELECT 'C',13,DATE '31-Jan-2017','Q'
UNION ALL SELECT 'B',2, DATE '31-Jan-2017','R'
UNION ALL SELECT 'C',3, DATE '28-Jan-2017','P'
UNION ALL SELECT 'D',23,DATE '29-Jan-2017','Q'
UNION ALL SELECT 'B',11,DATE '30-Jan-2017','R'
UNION ALL SELECT 'A',17,DATE '25-Jan-2017','P'
UNION ALL SELECT 'C',13,DATE '26-Jan-2017','Q'
UNION ALL SELECT 'D',17,DATE  '2-Feb-2017','R'
UNION ALL SELECT 'B',23,DATE  '3-Feb-2017','P'
UNION ALL SELECT 'A',45,DATE  '4-Feb-2017','Q'
UNION ALL SELECT 'B',3, DATE  '5-Feb-2017','R'
)
-- end of input , start real query here:
SELECT * FROM input
LIMIT 1 OVER(PARTITION BY Name ORDER BY payment_date)
;

Happy playing ... Marco the Sane



来源:https://stackoverflow.com/questions/44167796/find-the-first-key-by-date-field-using-sql-and-output-also-have-other-fields

易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!