using an aggregated field also as a filter

[#1]

hello,

i want to use an aggregated field as a filter as well.

i work at a company that sells an app for improving the install rate of applications (like hotels.com). some of the installs are invalid though due to technical issues.

i have a column that sums all the valid installs (sum(valid_installs)).

the field "valid installs" shows 0 for invalid install and 1 for valid install.

i want to be able to filter the data with this field, so to create a list box on the side that shows the values "1" and "0".

the problem is that in the loading process i have an aggregation on the field so i can't show it again in the Load clause.

we use amazon redshift database so the option of using partitions (like in oracle) is irrelevant.

would appreciate a solution to problem.

thanks.

[#2]

Hi Yaron,

if you are aggregating this field up in your script. You are loosing the more granular data regarding if a specific installation was valid or not and as you rightly say your field will no long be just 1 or 0.

My question is if you are only using that field for the expression sum(valid_installs), why the need to aggregate? If you don't aggregate the field you can still use that expression and then be able to bring the field in as a listbox too?

hope that helps

Joe

[#3]

indeed, your suggestion makes sense.

i wonder though, whether that wouldn't slow down the process.

currently the aggregations are built like this:

hourly aggregation upon which the daily aggregation is built upon which the weekly upon which the monthly.

i suppose cancelling this step, and making all the aggregation in the UI might cause the program to slow down, wouldn't it?