column group: deactivate formula eval

[hogi]

☐ cool new feature
☐ could help many users!

☐ removes a „bug“

☑ nice to have

☐ nobody needs it

my wish: 
possibility to deactivate Formula evaluation for all formulas in a column group via right click context menu of the column group.

more wishes submitted by  

[Sarah-Sylvestre]

@hogi - Thank you for your suggestion! We have captured your request and will take it under consideration.

[SamGardner]

@hogi thanks for another interesting suggestion.  After discussion with our developers, we would like to have more information on why this is important to you.  Can you describe or provide an example of how this would help you?  

[hogi]

This wish is related to 
https://community.jmp.com/t5/JMP-Wish-List/fast-way-to-enable-disable-formula-calculations/idi-p/547...

in the other wish I explained how users could benefit if there was an easy (one click) way  to enable/disable evaluation of a formula.

This wish is an add on:
If several column are grouped together in a column group, one could enable / disable / "trigger once" the evaluation of such formulas.

most important application case:
it's almost impossible to work with Jmp if the data set is large and if there are Formulas with selected()
But such formulas are just needed to highlight selected data points in the respective plots.
So: if such plots are not opened, one can disable such formulas and work much faster.

With the current version of Jmp, the user has to go to each individual Column formula and disable the respective formula.
Alternative approach: disable formula evaluation for the whole data table - but this would be like swimming without water.

Other applications:

• a group for formulas that have to be executed when new data is added to the data table.
  --> disable them if no data is added.
• a group for formulas using Excluded() - which just have to be updated if excluded is changed.
  

  At the moment, such Column formulas update after every change of the row selection (!!!)

  If there was an option to disable such a group of formula columns, one could work for some days with a very fast Jmp.
  Select rows, exclude rows etc. ...
  some weeks later one could enable the group again to use the columns again.

[hogi]

further details can be found here:
https://community.jmp.com/t5/JMP-Wish-List/fast-way-to-trigger-one-time-formula-evaluation-for-a-spe... 

[mia_stephens]

Hi @hogi , can you help us to better understand the underlying problem? There are a number of possible ways to make it easier to suppress formula evaluation. However, it sounds like the underlying problem is related to performance when evaluating formulas in large tables. Is this the case? Our development team has made several changes in JMP 17 and JMP 18 to improve performance with formula evaluation (in response to this and other related requests).

Questions from development: Have we made it fast enough in 18? If not, can you provide us with anonymized data that has performance issues with formula evaluations under your use case so we can test and continue to improve the performance?

Thank you!

[hogi]

Hi Mia,

the issue comes with formulas which contain Excluded() and Selected().
Such formulas are triggered every time the user changes the row selection:
Why does selecting/deselecting rows trigger column formulas to be re-evaluated? 

The problem was there for a while - wow, it got fixed in Jmp 18, right?
many, many thanks to Jmp development

dt = Open( "$SAMPLE_DATA/Airline Delays.jmp" );
wait(0);
dt << Concatenate(	dt,	dt,	dt,	Append to first table(1));

nr = NRows(dt);

// issue 1: formula eval gets triggered by EVERY row selection
add col = function({},
dt << New Column("x",
	Formula(
	If(row()==nr,Print(as date(today())));
	Col Mean( :Elapsed Time, :Airline, Excluded() ))
)
);

For (i=1,i<10,i++, addcol())

[hogi]

There are still places to improve, e.g.:
Col Aggregation with a string comparison in the If expression. 

Below you can find a sample script where Jmp formula evaluation is  NOT "fast enough".

(the second formula equation)

Interesting that changing the column type to compact doesn't help.

what's the difference between :Day of Month==1 and :Month == "Dec" for a compact column?

Just imagine a speed issue like this in combination with the  excluded() issue ...

dt = Open( "$SAMPLE_DATA/Airline Delays.jmp" );

dt:Month << Compact();

dt << New Column("x",
	Formula(
	Col Mean( If( :Day of Month == 1, :Elapsed Time, . ), :Airline ))
);

// who dares to enable this column? "compact" doesn't help 
/*
:Month << Delete Property( "List Check" ) << Compact();
dt << New Column("y",
	Formula(
	Col Mean( If( :Month == "Dec", :Elapsed Time, . ), :Airline))
)  

[hogi]

After removal of the rowstate excluded() bug, the need for this feature is reduced