Hi mansi.
When you load data from QVD:s you preferably do it optimized.
To do an optimized load, simply load the individual fields, or * without doing any operations or conditions.
This is the prefered way of loading data.
However, you cannot always do it since you need to do operations, for example date, num, count, where, exists et cetera. As soon as you do only one of these operations, you break the optimized load and the load is unoptimized.
Optimized load just means to transfer data from the QVD directly into RAM, hence it is optimized. It never needs to look at the data, it just transfers data into RAM.
If you instead do an operation, like date or num, the load needs do do the operation on each row inside the QVD before it is loaded into RAM.
Binary loads are even quicker than loading from QVD because it loads the whole QVW-file as a whole and copies the whole data model into RAM in one go. So if binary load can be used, it is very efficient.
I hope this will kick you in the right direction.
Cheers!
VeryOptimized:
BINARY MyFile.qvw;
Optimized:
LOAD
Field1
,Field2
;
Optimized:
LOAD *;
UnOptimized:
LOAD
num(Field1) as Field1
Field2
;
UnOptimized:
LOAD
Field1
,Field2
WHERE Field1 > 0
;
Kind regards
Magnus Åvitsland
BI Consultant
Framsteg Business Intelligence Corp.
