topic Re: Step by Step Row "list" building: Get All Rows, then Get Rows Where X, then Get Rows Where Y (and X) in Discussions

Step by Step Row "list" building: Get All Rows, then Get Rows Where X, then Get Rows Where Y (and X)

MeanChris — Thu, 06 Nov 2025 17:19:20 GMT

I'm trying to understand working with objects that are containers of rows. Even that description may be incorrect.

What is "FilteredRows" in this context?

dt = Current Data Table();
FilteredRows = dt << get rows where( :Date == 12May2025 );
Show( N Rows( FilteredRows ) );

To back up slightly, is there a JSL command to set an object to ALL rows in a datatable? Or is that never necessary?

The closest I can find that avoids 'Select All Rows' is:

AllRows = dt << get rows where(1);

Feels like cheating though.

More importantly, I'd like to be able to apply additional filters to a subset of rows

This does not work, but conveys the intent:

XFilteredRows = dt << get rows where( :X == 42 );
XYFilteredRows = XFilteredRows << get rows where(:Y ==1.21);

Yes, I can do both in one line

XYFiteredRows = dt << get rows where(:X==42 & :Y==1.21);

However I plan to use this in a For Each Row loop, where X removes 99% of the rows, and I'll need to subsequently filter more parameters independently, i.e. rows that meet XY, and rows that meet XZ and rows that meet X but not Y, etc...

I believe it will be faster on large tables to keep the X filtered rows.

On a related note, one result I'm trying to get is very similar to the 'Group By' result, but with a custom formula.

JSL of a column formula:

Col Mean( :Parameter, :X )

'Group By' is very convenient for created column formula from built in functions, but when I want to do more complex math I don't know how to achieve the same grouping effect.

Re: Step by Step Row "list" building: Get All Rows, then Get Rows Where X, then Get Rows Where Y (and X)

hogi — Fri, 07 Nov 2025 05:34:21 GMT

The "container" is a (1 dimensional) "matrix".

you can also create a matrix manually:

myRows= Matrix ({1,2,3})

myRows = 1::3

myRows = [1, 2, 3]

There are 2 ways to intersect (and other set operations like union, difference)

a) using associative arrays

aa1 = associative array (myRows1);
aa2 = associative array (myRows2);

aa1 << intersect(aa2);
// aa1 << remove(aa2);
list of rows = aa1 << get keys;
rows= matrix(list of rows);

b) [new in JMP19]: using lists

list1 = as list (myRows1);
list2 = as list (myRows2);

list of rows = set intersect(list1, list2);
list of rows = set intersection(list1, list2);
//list of rows = set difference(list1, list2);
rows= matrix(list of rows);

Intersections for Matrices? I guess: no

Re: Step by Step Row "list" building: Get All Rows, then Get Rows Where X, then Get Rows Where Y (and X)

MeanChris — Thu, 06 Nov 2025 23:15:44 GMT

hogi,

You've rescued me again. Thanks!

One clarification, I used your syntax

list of rows = set intersect(list1, list2);

initially and it turned blue so I thought it was okay, but odd that it didn't capitalize. Then I looked it up in scripting guide and realized it was "Set Intersection". Now I'm good to go. Cheers!