Scripted field & range analysis not working properly?

[SalvoDM91 (Salvatore Di Maria)]

Hi guys,

happy new year
I have a scripet field (number, percentage) with this code:

double tot_acc_nom_gruppo = doc['tot_acc_nom_gruppo'].value;
double tot_acc_pond_gruppo = doc['tot_acc_pond_gruppo'].value;
double delta_pratica = 1;
if((tot_acc_nom_gruppo != 0) && (tot_acc_pond_gruppo != 0)){
    delta_pratica = tot_acc_nom_gruppo / tot_acc_pond_gruppo;
}

return delta_pratica;

This painless code produces some results like:
47.345% or 137.238% and this is fine for me.

But if I try to create a table with a filter column in order to see the unique count of my documents for each range (0 - 10%, 10% - 20%, 20% - 30%) the total sum of unique count for each range is greather then the real unique count of my documents.

It's like some document is part of 2 or more range. Could you please help me?

Thanks!
Salvo

[Dzmitry]

Hi @SalvoDM91 ,

Could you share an example of your aggregation? It could be the same document ends in different buckets (data sets) based on your conditions.

Regards, Dima

[SalvoDM91 (Salvatore Di Maria)]

Ciao Dzmitry, I just found the problem.
I have more then 900K documents in my index and I see inside the documentation that Unique count is an an approximate count of distinct values.

This is a big issue for me... there is any way in Kibana in order to see the exact number?

Thanks in advance
Salvo

[Dzmitry]

I'm afraid with these amount of docs Elasticsearch might not return the exact count, but approximate one. But I would double check with Elasticsearch community by posting question there.

Regards, DIma