通過OPENXML解析XML時出錯

Question

我有一個55MB的XML，並且正在嘗試使用OPENXML將其切碎，因爲它看起來比正常的XML切分更快。 XML的結構是這樣的：

<DATA_EXPORT> 
<HEADER> 
    <RECDATE> 
     <START>2011-03-16</START> 
     <END>2012-02-10</END> 
    </RECDATE> 
    <SOME_COUNT>10288</SOME_COUNT> 
    <QUESTION_MAP> 
     <QUESTION> 
      <SERVICE>OU</SERVICE> 
      <VARNAME>UNIT</VARNAME> 
      <QUESTION_TEXT></QUESTION_TEXT> 
     </QUESTION> 
       250 more nodes like <QUESTION> 
    </QUESTION_MAP> 
</HEADER> 
<SOMENODES> 
    <SURVEY_ID>448817197</SURVEY_ID> 
    <CLIENT_ID>58</CLIENT_ID> 
    <SERVICE>OU</SERVICE> 
    <RECDATE>2011-03-29</RECDATE> 
    <DISDATE>2010-03-29</DISDATE> 
</SOMENODES> 
    : 
    : 
    1000s of nodes like <SOMENODES> 
</DATA_EXPORT> 

，我使用下面的查詢獲取數據出來

declare @xmlData varchar(max) = null 
     ,@iDoc int = null 
     ,@xml xml = null 

select @xmlData = cast(@xml as varchar(max)) 

    exec sp_xml_preparedocument @iDoc OUTPUT, @xmlData 

    select * 
      ,getdate() 
    from openxml(@iDoc, '//DATA_EXPORT/SOMEDATA', 2) 
    with (
     surveyId  varchar(50)  'SURVEY_ID[1]', 
     clientId  int   'CLIENT_ID[1]', 
     [service]  varchar(50)  'SERVICE[1]', 
     recieveDate  datetime  'RECDATE [1]', 
     dischargeDate datetime    'DISDATE [1]' 
    ) 
    option (optimize for unknown) 

我已經wrritten這是存儲過程和我得到以下錯誤的節點：

The XML parse error 0xc00ce562 occurred on line number 1, near the XML text "<". 
Msg 6602, Level 16, State 2, Procedure sp_xml_preparedocument, Line 1 
The error description is 'A declaration was not closed.'. 
Msg 8179, Level 16, State 5, Procedure q_ImportSurveyMasterDetails, Line 81 
Could not find prepared statement with handle 0. 
The statement has been terminated. 

任何人都可以幫助我什麼是行號錯誤，如錯誤所示？ 還有沒有更好更快的方式來粉碎一個55MB的XML文件？

在此先感謝。

Answer 1

問題＃1：The XML parse error 0xc00ce562 occurred on line number 1, near the XML text "<"。 通過將@xmlData更改爲@xml來修復，即不將其轉換爲varchar（max），只是將它作爲xml傳遞。感謝@dan radu爲您提供的所有幫助。

問題＃2：合併語句中的option關鍵字。它似乎不能在合併中使用，因爲合併語句本身就是一個單獨的sql語句。它應該在合併語句的末尾使用，例如

merge dbo.table1 as target 
using (
     select 
      n.value('(SURVEY_ID)[1]', 'bigint') as surveyId 
      ,n.value('(CLIENT_ID)[1]', 'int') as clientId 
      ,n.value('(SERVICE)[1]', 'varchar(50)') as [service] 
      ,n.value('(RECDATE)[1]', 'datetime') as recieveDate 
      ,n.value('(DISDATE)[1]', 'datetime') as dischargeDate 
     from @xml.nodes('//DATA_EXPORT/SOMENODES') x(n) 

     ) as source 
on target.surveyId = source.surveyId 
    and target.[service] = source.[service] 
when matched then 
update set 
     target.clientId = source.clientId, 
     target.[service] = source.[service], 
     target.surveyRecieveDate = source.recieveDate, 
     target.dischargeDate = source.dischargeDate 
when not matched then 
insert (surveyId, clientId, [service], surveyRecieveDate, dischargeDate) 
values (surveyId, clientId, [service], recieveDate, dischargeDate) 
option (optimize for unknown);