避免基于 PostgreSQL 中的多个列的重复条目

Question

我正在尝试从以下结果集中删除重复条目：

ID  Name    missionID   teamID  eventTimestamp  
25  C   1000415147500265653 1   02-11-2014 14:04  
25  C   1000415147500265653 1   02-11-2014 14:04  
25  C   1000415147500265653 1   02-11-2014 14:04  
25  C   1000415147500265653 1   02-11-2014 14:04  
25  C   1000415147500265653 2   02-11-2014 14:06  
25  C   1000415147500265653 2   02-11-2014 14:06  
25  C   1000415147500265653 2   02-11-2014 14:06  
25  C   1000415147500265653 2   02-11-2014 14:07  
25  C   1000415147500265653 2   02-11-2014 14:07  
25  C   1000415147500265653 1   02-11-2014 14:08  
25  C   1000415147500265653 1   02-11-2014 14:08  

对于 targetID、targetName、missionID、teamID 和 eventTimestamp 的每个组合，我只需要 1 个条目。所以我希望结果集看起来像这样而不是上面的：

targetID    targetName  missionID   teamID  eventTimestamp  
25  C   1000415147500265653 1   02-11-2014 14:04  
25  C   1000415147500265653 2   02-11-2014 14:06  
25  C   1000415147500265653 2   02-11-2014 14:07  
25  C   1000415147500265653 1   02-11-2014 14:08  

如您所见。可能有其他条目具有相同的 targetID、targetName、missionID、teamID 组合，但时间戳不同 - 我需要保留这些条目，因此我不能只在此运行最大或最小日期。我尝试过使用 distinct、ranking 和 group by，但似乎没有任何效果，我不断收到重复的条目。如果我在 date 上使用 max 或 min，我只会得到每个 targetID、targetName、missionID、teamID 的一个条目——这不是我想要的

我最初的查询只是为了得到重复的结果集是这样的：

select targetID, targetName, missionID, teamID, eventTimestamp from (
select a.characterID, a.targetID, a.targetName, a.missionID, b.teamID as teamID, a.eventTimestamp from events_live a
left join (
select distinct characterID, teamID from events_live where eventName = 'missionStarted' or eventName = 'characterCreated') b on a.characterID = b.characterID 

where eventName ='missionObjective' and a.eventTimestamp > '2014-10-31' and a.eventTimestamp < '2014-11-07' and objectiveName = 'Capture' and (targetID ='25' or targetID='26' or targetID ='27' or targetID = '28' or targetID = '29') and b.teamID is not null
order by missionID, targetID, eventTimestamp
)t1

order by missionID, targetID, eventTimestamp

我已经尝试在原始对象的基础上使用 group by targetID、targetName、missionID、teamID、eventTimestamp 进行另一次选择 - 但我仍然得到重复项。我试过在 eventTimestamp 上运行 distinct，但它仍然没有给我想要的

Answer 1

试试这个：

SELECT targetID, targetName, missionID, teamID, date_trunc('minute',eventTimestamp ) AS eventTimestamp
FROM (
    SELECT a.characterID, a.targetID, a.targetName, a.missionID, b.teamID as teamID, a.eventTimestamp 
    FROM events_live a
    LEFT JOIN (
        SELECT distinct characterID, teamID from events_live 
        WHERE eventName = 'missionStarted' 
        OR eventName = 'characterCreated') b 
    ON a.characterID = b.characterID     
    WHERE eventName ='missionObjective' 
    AND a.eventTimestamp > '2014-10-31' 
    AND a.eventTimestamp < '2014-11-07' 
    AND objectiveName = 'Capture' 
    AND (
        targetID ='25' 
        OR targetID='26' 
        OR targetID ='27' 
        OR targetID = '28' 
        OR targetID = '29') 
    AND b.teamID IS NOT NULL
    ORDER BY missionID, targetID, eventTimestamp) t1
GROUP BY targetID, targetName, missionID, teamID, eventTimestamp    
ORDER BY missionID, targetID, eventTimestamp

编辑： date_trunc('minute',eventTimestamp ) 应该可以解决问题

Answer 2

CREATE UNIQUE INDEX INDEX_NAME ON TABLE_NAME(COL1, COL2,...);

这将解决问题。