比较2条记录并在POSTGRESQL中使用YES或NO更新一个新字段

如何解决比较2条记录并在POSTGRESQL中使用YES或NO更新一个新字段

我有一个要求，我们需要在POSTGRESQL中使用“ for_loop”语句， 有人可以提供CODE帮助吗？

表名nex_outages具有字段：

• 序列号
• outage_type
• outage_start_date
• last_modified_date
• 中断事件名称。

我们必须按以下顺序排序：序列号，中断类型和中断开始日期。然后将序列号的第一行与第二行的序列号进行比较。

如果两者均匹配，则将第一行outage_type与第二行outage_type进行比较，如果两者均匹配，则检查第一行outage_start_date与第二行outage_start_date之间的差异，如果差异超过61天，则设置is_duplicate字段（新字段）='第一行为“否”，并且如果outage_start_date的差值小于或等于61天，则将第一行last_modified_date与第二行last_modified_date进行比较，并且最近一次设置的last_modified_date为is_duplicate ='NO'

如果第二行last_modified_date是最近的，则将NO分配给第二行，然后在第一行和第三行之间进行下一个以上比较，然后循环。

目前我的表有50,000条记录。寻找“ FOR_LOOP”查询来实现这一目标

样本数据

|-------------------------------------------------------------------------------------------------------------------------------|
| outage_type           | outage_event_name | outage_start_date     | serial_number | last_modified_date    | created_date
---------------------------------------------------------------------------------------------------------------------------------
| Major Inspection      | 83230             | 11/1/2033 0:00        | SY0073277     | 12/21/2017 8:31       | 12/21/2017 8:31   |
| Major Inspection      | 54578             | 11/1/2024 0:00        | SY0073277     | 12/23/2016 8:31       | 12/23/2016 8:31   |
| Major Inspection (MI) | 83231             | 10/30/2033 0:00       | SY0073277     | 4/12/2019 0:39        | 12/21/2017 8:31   |
| Major Inspection (MI) | 115317            | 11/1/2024 0:00        | SY0073277     | 11/8/2018 6:46        | 3/5/2018 14:58    |
| Major Inspection (MI) | 54579             | 2/3/2024 0:00         | SY0073277     | 4/11/2019 23:58       | 12/23/2016 8:31   |
| Minor Inspection      | 54576             | 12/27/2029 0:00       | SY0073277     | 12/23/2016 8:31       | 12/23/2016 8:31   |
| Minor Inspection      | 54577             | 12/26/2029 0:00       | SY0073277     | 4/12/2019 1:24        | 12/23/2016 8:31   |
| Minor Inspection      | 83229             | 11/26/2028 0:00       | SY0073277     | 4/12/2019 0:36        | 12/21/2017 8:31   |
| Minor Inspection      | 83228             | 11/21/2028 0:00       | SY0073277     | 12/21/2017 8:31       | 12/21/2017 8:31   |
| Minor Inspection      | 54575             | 4/24/2019 0:00        | SY0073277     | 5/25/2019 0:31        | 12/23/2016 8:31   |
| Minor Inspection      | 86703             | 4/20/2019 0:00        | SY0073277     | 4/12/2019 0:17        | 1/10/2018 19:46   |
| Minor Inspection      | 54574             | 4/20/2019 0:00        | SY0073277     | 1/10/2018 16:16       | 12/23/2016 8:31   |
|-------------------------------------------------------------------------------------------------------------------------------|

预期输出：

CREATE OR REPLACE FUNCTION duplicates() 
RETURNS VOID 
AS $$
DECLARE 
cur cursor for 
Select
outage_type,outage_event_name,outage_start_date,serial_number,last_modified_date,created_date,MIN(OUTAGE_START_DATE) OVER (partition by OUTAGE_TYPE,SERIAL_NUMBER ORDER BY OUTAGE_START_DATE desc 
ROWS BETWEEN 1 PRECEDING AND 1 PRECEDING) 
as nextrecord,lead(OUTAGE_START_DATE) OVER(partition by OUTAGE_TYPE,SERIAL_NUMBER ORDER BY OUTAGE_START_DATE 
desc) osd_lead,lead(last_modified_date) OVER(partition by OUTAGE_TYPE,SERIAL_NUMBER ORDER BY OUTAGE_START_DATE 
desc) lmd_lead,lag(OUTAGE_START_DATE) OVER(partition by OUTAGE_TYPE,SERIAL_NUMBER ORDER BY OUTAGE_START_DATE desc) 
 osd_lag,lag(last_modified_date) OVER(partition by OUTAGE_TYPE,SERIAL_NUMBER ORDER BY OUTAGE_START_DATE 
 desc) lmd_lag,lead(outage_type) OVER(partition by OUTAGE_TYPE,SERIAL_NUMBER ORDER BY OUTAGE_START_DATE desc) 
 otg_typ,'NO' as is_duplicate
 FROM sot_steam_da.nex_outages_src
 where serial_number = 'SY0073277'
 and outage_type = 'Minor Inspection'
 order by outage_start_date desc,last_modified_date desc,outage_type desc;

 cur_row sot_steam_da.nex_outages_src%rowtype;

 BEGIN
FOR cur_row in cur 

LOOP
    Update sot_steam_da.nex_outages_src 
    set is_duplicate = if (outage_start_date::date - osd_lead::date) <= 61 and outage_type = otg_typ THEN
                             'No';
                         elseif (last_modified_date <= lmd_lead) then 'Yes';
                         else 'No';
                              end if;
                               
    
END LOOP;
END;
$$ 
LANGUAGE plpgsql;

此代码未针对整个需求正确编码，并且上面的代码在第1条IF语句后引发语法错误，我们尝试对第1条IF记录使用上述代码中的where条件获得准确的输出。

解决方法

好的，您似乎在组合功能方面取得了合理的进步。关于UPDATE语句（尽管我不是PostgreSQL专家），我相信您的语法确实不正确。 IF子句在UPDATE / SET上下文中无效。您需要改用CASE语句，如下所示。希望这可以帮助您走得更远。

UPDATE sot_steam_da.nex_outages_src 
   SET is_duplicate = CASE WHEN (outage_start_date::date - osd_lead::date) <= 61 
                            AND outage_type = otg_typ 
                           THEN 'No'
                           WHEN (last_modified_date <= lmd_lead) 
                               THEN 'Yes'
                           ELSE     
                               'No'
                      END;