Is there efficient code that implements multiple columns of all non-null rows?

lwx228

I can only do it this way, if it's more than 100 columns it's too inefficient.

Thanks!

dt=Open("$SAMPLE_DATA\Missing Data Pattern.jmp");
dt << Select Where(
	abs(As Column(1))>=0&abs(As Column(2))>=0&abs(As Column(3))>=0&abs(As Column(4))>=0
);
Wait( 0 );
d3 = dt << Subset( Output Table( "text" ), Selected Rows( 1 ), selected columns( 0 ) );
Wait( 0 );

cwillden

You could do something like this:

dt = Current Data Table();

//Get list of current names:
cols = dt << Get Column Names();

//Build a string iteratively to select non-missing from each column
str = "dt << Select Where(!Is Missing(:Name(\!""||char(cols[1])||"\!")) ";
for(i = 2, i<= N Items(cols), i++,
	str = str||"& !Is Missing(:Name(\!""||char(cols[i])||"\!"))"
);
str = str ||")";

//parse and evaluate the string
eval(parse(str));