Discussion Board for collaboration on QlikView Scripting.
Hi Folks, I am trying to create one intermediate QVD and first I tried with Precedent Load:
SubField(A,'.',1) as X,
SubField(A,'.',2) as Y,
ApplyMap('MAP_T',AutoNumber(LEFT(A,LEN(A)-3)),'No') as FLAG
LOAD DISTINCT A,B,C,D FROM [$(vQVDpath\TEST_CHECK.qvd)] (qvd);
The TEST_CHECK.qvd have 4M rows
For this code I got 4M rows in FINAL_QVD table also. But I am expecting only 1M ditinct rows only.
Again, I re-write my script with Resiednt Load below:
DROP TABLE T;
With this, I got 1M rows in my new table FINAL_QVD.
I wonder, what is wrong with precedent load
What is the difference b/n Precedent Load & Resident Load
Put the distinct clause in the top load statement, and the preceding load will remove the duplicates. A distinct clause in the following load statements will be disregarded.
View solution in original post
Distinct statement is not working this way with Precedent LOAD.
Thanks for Reply HIC....
I wondor, Which load give better performance?
use Preceding Load ,this one better performance,but resident Load slow script and slow ur performance
I think you should make some tests, because it depends.
Just have made some tests with mapping and 6M rows. In my case "optimized QVD load+mapping from Resident" is faster than "loading from QVD and mapping with preceding load".
I guess it's because using preceding load here turns off optimized qvd load.