topic Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL in App Development

How Set Analysis works when measure field exists in different table and Dimension value is NULL

Digvijay_Singh — Sat, 16 Nov 2024 19:53:11 GMT

Hi,

I am trying to understand why set analysis drops some rows when chart dimension value is null and measure field exists in different than chart dimension field.

I have attached sample app, very similar to what we have in production app. The main question is, why left chart shows count 2 against null dimension but right one shows 1 against null dimension.

I observed if I replace set analysis expression with IF statement, both chart shows 2 against null dimension value.

Any help will be highly appreciated.

Thanks,

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

Ksrinivasan — Thu, 17 Oct 2019 02:36:16 GMT

hi,

can you send as QVW file,

K. Srinivasan

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

thi_pham — Thu, 17 Oct 2019 08:35:22 GMT

Interesting, I cannot understand why right chart return values like that, problem should be in the multi value of Pat ID key1 in Link_Table, I tried to make a left join between Link_Table and Patient_Table, then right chart will be the same of left one. Just for your information. Look forward to someone explain this case.

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

Digvijay_Singh — Thu, 17 Oct 2019 14:31:26 GMT

Attached is the QVW reflecting the same issue.

Thanks for your time.

DJ,

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

Digvijay_Singh — Thu, 17 Oct 2019 15:41:43 GMT

Thanks for your response.

Its really very confusing case, I cannot join link table and patient dim table but I would certainly will see how that is making difference. I have been trying to understand how 'IF' inside count in the right chart makes both chart match, so not sure if its something related to table associations? or its linked to the behavior of set analysis when fields come from different tables.

Thanks,

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

Ksrinivasan — Fri, 18 Oct 2019 03:10:16 GMT

Hi,

please refer below table and chart, it will make clear you,

K. Srinivasan.

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

thi_pham — Fri, 18 Oct 2019 03:29:57 GMT

It doesn't make the problem clearer with me. Just add 1 more dimension (_f2f_enc_ind or _addl_visit_vol_enc_ind or _appt_complete_ind) to right chart , you will see the strange thing.

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

Ksrinivasan — Fri, 18 Oct 2019 05:01:33 GMT

Hi,

See below Table and chart, as you said dimension added,

K. Srinivasan

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

Digvijay_Singh — Fri, 18 Oct 2019 14:35:17 GMT

Thanks for your response. I know what is happening but I don't know why one null dimension row is dropped when measure field comes from other table.

DJ,

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

Digvijay_Singh — Fri, 18 Oct 2019 14:39:28 GMT

Another angle of the problem is - if in the right chart I use 'IF' instead of set analysis it start matching with the left chart.( I have kept 'IF' statement commented in the right chart.)

So not sure whats the root cause behind the difference. Its certainly a combination of set analysis behavior and table association.(Not just having measure field in another table)

Thanks,

Re: How Set Analysis works when measure field exists in different table and Dimension value is NULL

Digvijay_Singh — Fri, 18 Oct 2019 17:21:56 GMT

When I drop 'Key' column from the table, It drops the middle record. Strange, Even 'Key' removal wasn't making duplicate records situations in the table.