Sql 用现有数据对新数据进行分区

我正在尝试更新已添加到表中的新记录的索引字段。有些记录已经有索引，如何用下一个值序列更新新记录

Sample data:
    Index   INCIDENT_DT Incident_type   Event       CITY            COUNTRY
    1       1/2/16 0:00 Accident        Truck       Boston          NULL
    2       1/2/16 0:00 Accident        Car         Boston          NULL
    3       1/2/16 0:00 Accident        Ground      Sutton          USA
            1/2/16 0:00 Accident        NULL        NULL            NULL
            1/2/16 0:00 Accident        NULL        Chicago         NULL

SELECT  ROW_NUMBER() OVER (partition BY  INCIDENT_DT order by  INCIDENT_DT) AS Index, INCIDENT_DT  
FROM  (select distinct  INCIDENT_DT, Incident_type, [event], INCIDENT_CITY, INCIDENT_COUNTRY from input ) r

感谢您的帮助。

她的一个方法是联合：

---

约格什：谢谢你的主意。从来没有想过工会

这起作用了

SELECT  ROW_NUMBER() OVER (partition BY  INCIDENT_DT order by  grp, index, INCIDENT_DT) AS grp, index, INCIDENT_DT  
FROM  (select distinct 1 as grp, index, INCIDENT_DT, Incident_type, [event], INCIDENT_CITY, INCIDENT_COUNTRY from input where index is not null
union
select distinct 2 as grp, index, INCIDENT_DT, Incident_type, [event], INCIDENT_CITY, INCIDENT_COUNTRY from input where index is null
 ) r

---

使用UPDATE函数。我将停止手动执行此操作，并将属性应用于索引字段。我希望有一种方法可以在上面的评论中给@scsimon更多类似的东西。这个过程是精神错乱的人格化。这是一种竞争条件，极易发生错误和/或冲突。抱歉，已使用标识列，更新将不起作用。我正在尝试更新组上的索引你是什么意思？下面是一个你需要做的例子。

Index   INCIDENT_DT Incident_type   Event   CITY    COUNTRY
1       1/2/16 0:00 Accident        Truck   Boston  NULL
2       1/2/16 0:00 Accident        Car     Boston  NULL
3       1/2/16 0:00 Accident        Ground  Sutton  USA 
4       1/2/16 0:00 Accident        NULL    NULL    NULL
5       1/2/16 0:00 Accident        NULL    Chicago NULL

SELECT  ROW_NUMBER() OVER (partition BY  INCIDENT_DT order by  grp, index, INCIDENT_DT) AS grp, index, INCIDENT_DT  
FROM  (select distinct 1 as grp, index, INCIDENT_DT, Incident_type, [event], INCIDENT_CITY, INCIDENT_COUNTRY from input where index is not null
union
select distinct 2 as grp, index, INCIDENT_DT, Incident_type, [event], INCIDENT_CITY, INCIDENT_COUNTRY from input where index is null
 ) r