Incremenatl Load logic :
I Have a large source : let's say 20 millions line :
1) Let's import this and store it in a qvd:
load * from source;
store Table into table.qvd(qvd);
2) Now, that we store these 20 millions in the qvd; we comment the previous script;
3) the next day, we have 21 millions lines in our source; instead of importing the whole 21 millions from our database:
a) Import the 20 millions already stored in the qvd
b) from the source database, we import only the new 1 million lines
Load * from Table.qvd(qvd);
load * from source where not exists (UniqueID); we just import the lines that do not exist in the QVD.
Hope this was clear?