Alternative for concatenation to flag data?

[Anonymous]

I have a csv file about 480 MB. It has 8 columns. column 5 to 8 are data columns which i concatenate one by one in order to flag the kind of data it has. This 4 tyimes concatenation is causing huge perf issue. Even I do the alternative of Crosstable, same problem. Please advice any other alternative.

[MK_QSL · Accepted solution]

Temp:

Load [Country Id],[Mfr ID], Period, Company, Data1, Data2, Data3, Data4 From YourQVDName;

Final:
Load [Country Id],[Mfr ID], Period, Company, 'Abs' as Flag, Data1 as Value Resident Temp;

Load [Country Id],[Mfr ID], Period, Company, 'Exc' as Flag, Data2 as Value Resident Temp;

Load [Country Id],[Mfr ID], Period, Company, 'Res' as Flag, Data3 as Value Resident Temp;

Load [Country Id],[Mfr ID], Period, Company, 'USD' as Flag, Data4 as Value Resident Temp;

Drop Table Temp;

=============================================

First load full data from QVD rather than loading 4 different times. Now use Resident load which is much faster than Optimized Load... This will definitely reduce your load time...

[Anonymous]

Thank you. Yes it did reduce my load time! It would be helpful if you can also guide how to reduce the size of the table. As my primary concern is sizeas it increases my QVw size a lot and make the data modeling slower. Thank you.

[MK_QSL]

There is no way to reduce the size as you want to load all data and also want to transform the table with different flags...

[Anonymous]

Yeah.. okay.. Thanks a lot!!