merge_ui

[cicdguy]

We use to call data_merge_module in the server but it doesn't return any information to ui. We might need to present ANL details in the encoding panel.

The reason for this is that left_join, full_join causes some confusion. Reportedly it's not clear what is the output of used datasets - even if we have datasets: ADSL, ADLB, ... in the encoding panel and datasets sizes in the flter panel - it's still vague what is the result of these joins.

merge_ui (or other name) can be put (optionaly) in the ui and can receive some information from the module.

Provenance:

Creator: gogonzo

[cicdguy]

NEST/teal.modules.general/issues/902 it can be quite confusing exactly what the analysis data set is:

Dataset 1:
Name   Age   Colour     
Fred    45    red           
Joy     56    green          
Nik     33    blue              


Dataset 2:
Name Age Result    Result_Date   
Fred   45   Good     1st Jan
Joy    56    Bad     1st Jan
Nik    33    Bad     1st Jan
Fred   45    Bad     1st Feb
Joy    56    Bad     1st Feb


Set outlier_var = Age from Dataset 1

If no categorical_var then ANL =
    Age 
    45
    56
    33

If categorical_var = Colour then ANL = 
Age Colour
45   red
56   green
33   blue

If categorical_var = Result from dataset 2, ANL = 
Age Result
45   Good
45   Bad
56   Bad
56   Bad
33   Bad

So we've changed the set of points we are trying to look for outliers 
in by choosing categories (is this ok?) but users do not get any information that
this is the case.

Provenance:

Creator: burkoffn

[cicdguy]

Do we know what we need here? If not then we probably need some discussion label here.

Is this what you are looking for:

Provenance:

Creator: pawelru

[cicdguy]

@pawelru yes, but I'm thinking weather it should be a in:

• filter panel
• data extract panel - along with Datasets information

Provenance:

Creator: gogonzo

[cicdguy]

For me (left) data extract panel is only about input specification. It's already too much there so I wouldn't put a new item in there. But let's discuss it.

Moreover, there are modules that has more than one analysis dataset (thus: how to name them? would it be clear for the end-user?). There are modules that has no analysis dataset per se (see missing data module). This should be also taken into account.

Provenance:

Creator: pawelru

[cicdguy]

I see both sides of the discussion here. On one hand, it's confusing, on the other hand, we already provide the code for the joins, so it should be clear what teal does.

Provenance:

Creator: kpagacz