Hi, I'm trying to load a .txt file with 350.000 lines and 200 columns!!, but if I load * and then store into a qvd, I lost 100.000 lines.
I tried importing the txt file into an access file forcing the column format into a text format and then loading this .mdb file in qlikview and works just fine (it loads the 350.000 lines)
I think the problem is in the way that the ODBC loads the .txt.
Is there a way to force the "text" format in qlikview ??
What I suspect you need to do is load the values as Text() from the text file and then store into the QVD. If trying to load using text() from the QVD it's too late as the values have already been stored as identical.
The reason for this is that QlikView tries to interpret the values numerically when possible because of performance reasons and distinct values are stored while recurring values are just stored as references to an earlier value keeping the size of the internal database well compressed.
So a list of values like this:
005, 05, 5 would all be stored as 005 with the value 5. If I load them as text() however they all keep their formatting and are stored in the internal table as distinct values.
It sounds like delimiter for "end of record" is missing or different for the 100.000 lines that do not load. Can you tell if the data from those missing lines is added to the end of other fields?
Considering what you said, I ve tried to load lest columns to control the end delimiters, but I think that qlikview transform the record.
I have to different products, one with code nº 'p00001' and another with code nº 'p000001' and when store into qvd, qlikview use 'p000001' for both of them.
Ive tried with de formula text(CodProd) but, still load 'p000001' for both.
Do you know how to force QV to load text in a field?
What I suspect you need to do is load the values as Text() from the text file and then store into the QVD. If trying to load using text() from the QVD it's too late as the values have already been stored as identical.
The reason for this is that QlikView tries to interpret the values numerically when possible because of performance reasons and distinct values are stored while recurring values are just stored as references to an earlier value keeping the size of the internal database well compressed.
So a list of values like this:
005, 05, 5 would all be stored as 005 with the value 5. If I load them as text() however they all keep their formatting and are stored in the internal table as distinct values.
