topic Re: Evaluating expression at the row level in QlikView

Evaluating expression at the row level

sampada0810 — Tue, 16 May 2017 13:41:43 GMT

Hi All,

I have a scenario where I am having a straight table with 2 dimensions and an expression.

Please find the screenshot below:

Dimension B is having 2 values for every value in dimension A.

And here I want to handle confidentiality, that is for particular selections the data cannot be shown, it will be NA.

So for my particular selection of Dimension A, if my confidentiality is failing for 1 value in dimension B, it should show NA for both the values of Dimension B. Because if I add a total to that column, from total anyone can get the value for NA.

Any idea how can I handle confidentiality at the row level.

Let me know if some more information or clarification required.

Re: Evaluating expression at the row level

Anonymous — Tue, 16 May 2017 13:46:11 GMT

Hi Sampada,

Use set analysis, depending on how you want it to show in the table depends on the aggregation you use (max, min, sum, count)

Try this:

Sum({$<TotalSales-={'NA'}>}TotalSales)

hope this helps

Re: Evaluating expression at the row level

sampada0810 — Wed, 17 May 2017 06:19:52 GMT

Hi Matthew,

If I use aggregation, it will display only 1 value.

But I want to show both the values, if any one is NA, both should be displayed as NA.

Re: Evaluating expression at the row level

marcus_sommer — Wed, 17 May 2017 07:10:56 GMT

The normal and common way to handle confidential data is to implement Section Access within the application which just removed the confidential data on a user-level by opening the application. But you will need some kind of a multi-staging process if you want to keep totals for fields which are data-reduced for users. For this requirement might be this posting helpful: Mask or de-identify data for certain users using Section Access.

Another approach which could be quite expensive would be to try o implement it by chart-expressions which query the user and if your show-conditions are true and which chart-level is beeing used - I mean something like:

if(match(osuser(), 'User1, 'User2') and maxstring(Field) adn dimensionality() > 1, 'then', else')

whereby you might need to nest your conditions quite heavily.

- Marcus

Re: Evaluating expression at the row level

Anonymous — Wed, 17 May 2017 07:29:29 GMT

I see, can you provide some sample data? It doesn't have to be real, just some fake data to get an idea of how it works, perhaps in an excel sheet.

Marcus has also made a good point below about Section Access