Excellent.  Thanks Larry and Karen. 



Karen, 

 The only reason I ask about the multi-table insert question is that I use MDI 
forms a lot.  

There is a very remote possibility that a user could insert a header record, go 
to another 

form and possibly insert a new record, and then come back to the first form and 
insert 

the detail records.  An example would be entering a sales record header, then 
adding 

a new inventory item and then coming back to enter the newly entered item on 
the sales 

order detail line.   



Truthfully, I do not believe I have this situation, but I am always  trying to 
"conceive" any 

possible bug or crash that could happen.  It also sometimes helps my 
understanding of 

the overall process. 



Thanks again everyone . 



-Bob 



----- Original Message ----- 
From: "Lawrence Lustig " < larrylustig @yahoo.com> 
To: "RBASE-L Mailing List" < rbase -l@ rbase .com> 
Sent: Thursday, January 28, 2010 2:22:01 PM GMT -06:00 US/Canada Central 
Subject: [RBASE-L] - Re: where COUNT=INSERT 



<< 
2)Does this use an index?  






3)The speed to retrieve the value will need to be very fast and this will 

be running on a network. 

>> 




CAUTION: This is speculation on my part, but it's the way I've always 
understood WHERE COUNT = INSERT to work. 




I believe that a pointer to the most recent INSERTed row is actually stored in 
R:Base's memory at the time of insert.  So when you retrieve the row, you're 
going directly to the correct address in the datafile , which would be even 
faster than using an index. 




And yes, R:Base can remember the last inserted row for each table separately. 

-- 

Larry 



Reply via email to