Filter by unstored calc

[winklestork]

I am struggling with something I think should be easy. I want to display in a portal related records who's parent record has a particular status. No records are displayed. See example.

[Søren Dyhr]

Read this:

http://www.fmforums.com/forum/showtopic.php?tid/191906/post/275767/#275767

--sd

[winklestork]

My schema is very similar to Untagged, which you directed me to. I have a calculation similar to cOpenJobID except the task_status is 0 or 1 and the job_status is the sum -- 0 meaning everything is complete. I have the features represented by the template. What the template lacks which I need are two portals that display tasks, independent of client and job, (1) that shows all incomplete tasks (because the person doing the task is not driven by client or job -- this is easy) and (2) which shows all complete tasks for incomplete jobs (because before a job is complete you can edit completed tasks). It's the "incomplete jobs" filter I can't figure out.

[Søren Dyhr]

Unfortunately have your upload gone wonky!

I am struggling with something I think should be easy. I want to display in a portal related records who's parent record has a particular status. No records are displayed. See example.

Attachment: filterbystatus.fp7.zip (9.23 KB) 21 View(s)

Couldn't you give it another stab and re-upload? Measures seems now to be taken by Steven...

Anyways, acting upon this:

I need are two portals that display tasks, independent of client and job,

Could it quite simply be done this way, as shown in attachment!

--sd

UntaggedMOD.zip

[winklestork]

Thank you for your time. The enclosed file (hopefully enclosed correctly this time) shows exactly what I'm trying to do and what doesn't work. Look at the 'Tasks by Status' layout.

UntaggedMOD2.zip

[comment]

Try it this way?

Untagged3.fp7.zip

[Søren Dyhr]

Why two globals?

--sd

[comment]

Because they were already there, and I was concentrating on the relationships. But you're right.

[winklestork]

THANK YOU. Soren and Comment, your help is very much appreciated.

I learned about GetAsBoolean, very useful.

The important lesson, which is probably second nature to someone more into relational databases, is to remember that a null value for a key prunes records when used in a relationship.

I used two globals because I thought an equal relation was an order of magnitude more efficient than a not equal relation for a large dataset in that an index can be traversed for the former and not the latter. (Right?). The task table grows by 6000 records per month.

You won't believe I spent hours and hours trying to figure this out, although I did something similar in another part of the application without actually understanding how it worked. Thanks again for your selfless, willing and generous help.