Get a list of the unique values in a column, matrix, or list

[#1]

Problem

Problem 1:  I need to get a list of the unique values in a data table column.

Problem 2:  I need to get a list of the unique values in a JSL data structure (list or matrix)

Solution

For problem 1:

Solution 1:  Use summary tables.  The JSL below shows how to use Summary() to get a table that will have a row with each unique value in a single column.   You can then do further JSL table commands to get those values from the summary table. 

_dt = Open( "$sample_data/big class.jmp" );
//create a summary table of the :sex column
_dt1 = _dt << Summary(
	Group( :sex ),
	Freq( "None" ),
	Weight( "None" )
);
//create a summary table of the :age column
_dt2 = _dt << Summary(
	Group( :age ),
	Freq( "None" ),
	Weight( "None" )
);

Problem 2:  

Now,  instead of a data table you have a JMP list or matrix:

sexlist = {"F", "M", "M", "F", "F", "M", "M", "F", "M"};
agelist = [14, 11, 12, 13, 15, 17, 15, 11, 12];

the complicated ways to do this would be:

a) put the values in the list into a new data table column, and then use the summary approach, OR

b) write a somewhat complicated for loop to search for and record unique values

Fortunately, there is an easier way, using associative arrays.  

Associative Array(sexlist) << Get Keys;
Associative Array(agelist) << Get Keys;

returns {"F", "M"} and {11, 12, 13, 14, 15, 17}, respectively (note that these are nice sorted list!).  If you would like the numeric values as a sorted vector, you can use

Matrix(Associative Array(agelist) << Get Keys);

which returns [11, 12, 13, 14, 15, 17].  

This approach can also be used in place of Solution 1 for Problem 1:

Problem1, Solution 2:

Associative Array(Column(_dt, "sex") << Get Values) << Get Keys;Matrix(Associative Array(Column(_dt, "age")<< Get Values)<<Get Keys);

returns {"F", "M"} and [11, 12, 13, 14, 15, 17]., respectively. What this is doing is 

• get the values from the specified column as a list (if the column is character) or a matrix (if the column is numeric)
• turn that list or martrix into an associative array
• get the keys for the associative array, which in this case are the unique levels of the values in the column, returning that as a list
• (if Matrix() is used on the resulting list, it will only work if the values in the list are all numbers).

[HP_Lau]

Summarize() is way much more efficient than associative array, it is very obvious with 2 million rows of data (1sec vs 8sec time lapse). If performance is critical, use Summarize() whenever possible.

Summarize( dt, smrz1 = by( :sex ) ); Summarize( dt, smrz2 = by( :age ) );

* Summarize() does not create intermediate data table as << Summary

Associative Array is great in the way that it saves the effort to put the list/matrix into data table, this is perfect solution for small data size.