topic Re: count difference in QlikView

count difference

ajaykumar1 — Sun, 16 Feb 2014 15:09:43 GMT

Hi All,

I have a small doubt.

The records count of a a table after reload and the count in the table box is showing different count.

But if am saving that file in csv after reload its showing exact count.I dont have the access to share the data.

What might be the possible reasons to get difference in table box.

Thanks,

Ajay

Re: count difference

rbecher — Sun, 16 Feb 2014 15:23:10 GMT

Hi Ajay,

a table box only shows distinct records (like a listbox shows distinct fields). So, the record count of the table box could be less then the record count of the table (loaded records).

- Ralf

Re: count difference

Sun, 16 Feb 2014 16:36:04 GMT

Hi,

for validation purposes you get full control, if you generate a systemid field inside your load table statement. Something like: Load rowno() as systemid, ....

Then use this systemid inside your tablebox.

Best

Sacho

Re: count difference

rbecher — Sun, 16 Feb 2014 16:42:44 GMT

..this would help but those kind of IDs (esp. for facts) are very memory consuming.

Re: count difference

Sun, 16 Feb 2014 16:56:00 GMT

Hi Ralf,

sequentiell numeric non key fields are in context of column based memory structures peanuts, especially with high cardinality fact tables.

Re: count difference

rbecher — Sun, 16 Feb 2014 17:03:29 GMT

No, you're wrong. There is no "sequentiell numeric" data type in QlikView. It's either a 4 or 8 byte (int/double) value for each number (plus some extra byte). It also blows the record with an additional pointer. And believe me, this kind of fields is what you should omit, if possible.