Formula for Number of Unique Categories in Column

[#1]

I am looking for the column formula to provide me with the count of unique categories in a column (maybe based on another columns categories). For instance, if a column has

A

A

B

B

B

C

then I would like to get back the value 3.

Thanks

[#2]

I think I remember the Summation formula being used for this, but my memory fails me.

[#3 · Accepted solution]

Try this formula.......of course, swap out the :Flow for the column you want to count the categories for

Summarize( x = by( :FLOW ) );
N Items( x );

If the data table is really large, you may want to use this more efficient version

If( Row() == 1,
	Summarize( x = by( :Flow ) );
	y = N Items( x );
);
y;

[#4]

This is PERFECT! Thank you Jim!

[#5]

I see how this would work for a script to create a column, but I've created a datatable that I update from a database and I'd like to have a column in that table with a formula to calculate the number of unique categories. Is there a way to do that directly in a column formula instead of though a script?

I found a hack using using Def() and Col Sum but it only works because i'm looking at a sorted field that's a unique id with numbrers. It doesn't work for actual text variables. Is there a better solution for that?

Here is the hacked formula:

Col Sum( Dif( :customer_id ) > 0 )

Thank you,

Aron

[#6]

The stated formula from my last response should do what you want.  

If( Row() == 1,
	Summarize( x = by( :Customer_ID) );
	y = N Items( x );
);
y;

Just cut and paste this into the formula editor

[#7]

What I always forget, is that a column formula can actually be an entire mini script; with many semicolons and lines of code.

[#8]

That worked great! I was thrown off by the red underline, I thought that meant it wasn't going to work.

I'm not actually sure I've ever tried pasting a formula directly in the column before, I thought I had to rebuild it using the explicit formula column functions... this is a huge improvement. Thank you so much for the quick reply!

Aron

[#9]

Mr. Nelson,

Can you describe in detail what this function is doing? Not sure I understand the "by" function.

Thanks,

Martin

[#10]

"By" is a fairly global concept in JMP.  If you run the Distribution Platform, and specify a "By" variable it will generate a separate Distribution for each of the different values of the "By" columns.  The same is for the Summarize() function.  Here is a quote from the JSL Syntax Reference for the Summarize() function.

     Help==>Books==>JSL Syntax Reference

"Using a BY variable changes the output from single values for each statistic to a list of values for each group in the BY variable."

Good documentation along with an example can also be fount in the Scripting Index

     Help==>Scripting Index==>Functions==>Statistical==>Sumarize

or

     Help==>Scripting Index.........search on Summarize