945769_en-US

liliconcom

What is criteria of column placement when sorting columns by data type?

I have table of 172 columns in a specific order that are a mix of character and nominal data types. Using the window on the left side of the table, I sort the columns by type which occurs ok. However, within each type grouping, the columns are no longer in a specific order but seem to have been placed randomly. What is the criteria for the column placements when the sort is done?

Data Blending and CleanupRe: What is criteria of column placement when sorting columns by data type?

According to 

Using JMP > Enter and Edit Your Data > Organize Data in Data Tables > Rearrange Columns in Data Tables 

Reorder By Data Type

Arranges the columns from left to right in alphabetic order by data type (expression, row state, character, or numeric).

So I would guess the columns are moved to data type groups (character, expression, numeric, row state) and stay then they stay in original order within those groups.

Re: What is criteria of column placement when sorting columns by data type?

Sadly, I don't want them arranged alphabetically which is the way JMP appears to want to arrange them, requiring more work at my end.. They are from survey and I want them kept in order seen. 

Re: What is criteria of column placement when sorting columns by data type?

You might impose the order you want with a prefix to the column name, such as an index value (01, 02, etc.). The ordered sorted by JMP would then be the same as the order by desired you.

Re: What is criteria of column placement when sorting columns by data type?

Here's a function that I believe does what you want.  Tested with JMP 19.04.  It breaks without the Eval( Eval Expr()).

arrange_columns = Function( {dt},
	num_columns = dt << Get Column Names( Numeric, String );
	char_columns = dt << Get Column Names( Character, String );
	ex_columns = dt << Get Column Names( Expression, String );
	rs_columns = dt << Get Column Names( RowState, String );
	dt << Reorder By Data Type();
	For Each( {group, index}, {num_columns, char_columns, ex_columns, rs_columns},
		If( N Items( group ) > 1,
			Eval( Eval Expr( dt << Move Selected Columns( Expr( group[2 :: N Items( group )] ), After( group[1] ) ) ) )
		)
	);
);