topic Evaluating Col Mean with Iterations in Discussions

Evaluating Col Mean with Iterations

Ksrzg01 — Thu, 19 Apr 2018 18:19:29 GMT

Hi everyone,

I'm having trouble evaluating a col mean function. As you can see below, the col mean(:age, :list_col_names[i]] doesnt insert the item in the list on the data table. With this current configuration, when I open the formula on the data table I get "Col Mean( :age, list_col_names[i] )" which doesnt correspond to anything on the list_col_names list.

list_col_names = fdt << getColumnNames( "String" );

For( j = N Items( list_col_names ), j > 0, j--,
	If( Left( list_col_names[j], 2 ) != "10",
		Remove From( list_col_names, j )
	)
);

For( i = 1, i <= N Items( list_col_names ), i++,
 fdt << New Column( list_col_names[i] || "_age", Numeric, Formula( Col Mean( :age, :list_col_names[i] ) ) ) 
);

on in an iteration

Re: Evaluating Col Mean with Iterations

vince_faller — Thu, 19 Apr 2018 23:00:03 GMT

you can't do :list_col_names[i], You're explicitly stating that you have a column called "list_col_names" and you're pulling the ith row of it.

Try this

Names Default to here(1);
fdt = open("$SAMPLE_DATA\Big Class.jmp");
list_col_names = {"sex", "name"};

For( i = 1, i <= N Items( list_col_names ), i++,
	//this is just it doesn't keep the variable 
	//in the table because i will change and then your column name will change
	Eval(
		EvalExpr(
			fdt << New Column( list_col_names[i] || "_age", 
				Numeric, 
				Formula( Col Mean( :age, 
					//replaces the string column name with an actual column
					Expr(Column(fdt, list_col_names[i]))
						
				) ) 
			)
		)
	)
);

Re: Evaluating Col Mean with Iterations

Ksrzg01 — Fri, 20 Apr 2018 14:38:52 GMT

Works great. Thank you Vince!