db2数据库上的sql查询

我正在使用此查询：

select substr("Message_Time",6,2) || '/' || substr("Message_Time",4,2) || '/' || '20'|| substr("Message_Time",2,2) || substr("Message_Time",8,2) || ':' || substr("Message_Time",10,2) as "Date",
       count(*) as "Fault", "Message_Location", "Service_Name_U", "Operation_Name_U", "Port_Namespace_U", 
      "Error_Code_U", "Error_SubCode_U", "Fault_Code_U", "Fault_String_U",
       "Requester_Identity", "Application_ServerName_U"
from "Fault_Log_Table_610"
where "Message_Time" >= 1181016220000000 and "Message_Time" < 1181017220000000 and
      "Operation_Name_U" = 'getDomandeDisabile'
group by substr("Message_Time", 6, 2) || '/' || substr("Message_Time", 4, 2)  || '/' || '20'|| substr("Message_Time", 2, 2) || substr("Message_Time",8,2) || ':' || substr("Message_Time",10,2), 
       "Service_Name_U", "Operation_Name_U", "Error_Code_U", "Error_SubCode_U", "Message_Location", "Fault_Code_U", "Fault_String_U",
       "Port_Namespace_U", "Requester_Identity", "Application_ServerName_U"

我需要添加2小时到日期字段，换句话说，我需要将数字2添加到substrMessage_Time，8,2

日期的格式为1181020164532000，其中：

第一个数字是世纪 接下来的两个数字是年份 接下来的两个数字是月份 接下来的两个数字是天数 接下来的两个数字是小时数 接下来的两个数字是分钟

---

最后三个数字是毫秒，我将预先计算时间戳解析，然后在CTE公共表表达式中添加2小时

然后我将在您想要的查询中使用这个即用数据。比如：

with x as (
select
    timestampadd(8, 2, -- add 2 hours
      timestamp_format( -- parse the VARCHAR into a TIMESTAMP
                       '20' || substr("Message_Time",2,15) || '000',
                       'YYYYMMDDHHMISSNNNNNN')
    ) as "Date",
    "Message_Location", "Service_Name_U", 
    "Operation_Name_U", "Port_Namespace_U", 
    "Error_Code_U", "Error_SubCode_U", "Fault_Code_U", "Fault_String_U",
    "Requester_Identity", "Application_ServerName_U"
from "Fault_Log_Table_610"
where "Message_Time" >= 1181016220000000
  and "Message_Time" < 1181017220000000
  and "Operation_Name_U" = 'getDomandeDisabile'
)
select
    "Date",
    count(*) as "Fault",
    "Message_Location", "Service_Name_U", 
    "Operation_Name_U", "Port_Namespace_U", 
    "Error_Code_U", "Error_SubCode_U", "Fault_Code_U", "Fault_String_U",
    "Requester_Identity", "Application_ServerName_U"
from x
group by "Date", 
    "Service_Name_U", "Operation_Name_U", "Error_Code_U", "Error_SubCode_U",
    "Message_Location", "Fault_Code_U", "Fault_String_U",
    "Port_Namespace_U", "Requester_Identity", "Application_ServerName_U"

---

将时间戳存储为时间戳值。不要使用VARCHAR。如果你使用VARCHAR，你会遇到很多问题。所以它会分解为：1-18-10-20-16-45-32-000？Century 1？或在表上创建一个视图，该视图将转换为时间戳，前提是首先要将列更改为时间戳需要花费太多的精力。您甚至可以让它生成一个列，或一个函数索引来帮助查询performance@PaulVernon我同意。对于这种情况，生成的列将非常有效。