could you please detail how you decide which values / fields are modified? Do you have a modification date in your table or something like this? And against what value do you compare (e.g. do you have a LastDate or something like this)? Then you could probably use a WHERE ModDate >= LastDate statement in your load.
If you want to perform an incremental load, please look also at the example in the QV cookbook, download available here:
Thank you Swuehl fot youre answer.
I read about the incremental load, but I don't thik that this statement can be used in my problem, because I have aprox. 500 tables to load that are modified frequently.
This means that I have to add an extra ModDate field to each table and after that to apply the condition WHERE ModDate >= LastDate.
I thought that there is a solution; when loading, to verify what differences are between the new qvd and the old loaded one.
Any idea will help me.
I managed to get something like this working a little while ago - give it a go. When you create the QVD files, combine the fields that are likely to change and wrap in a Hash() function (128, 160 or 256 bit depending on the number of records). Then use WHERE NOT EXISTS (HashField) in the load.
Hope this helps,
Youre method can be applied also if I'm interested on updated,deleted and inserted rows?
The Hash() function attributes a dimension to some specified fields and after the load statement, QV loads only the fields that have a different dimension from the previous load.
Is it ok what I have understood?