13381_en-US

liliconcom

Formula for grouped statistics<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />

Hi

I think it's trivial but I've only started to learn JSL syntax and I'm struggling with this one.

Saying I have a column "myValues" that contains my numeric values and a column "myGroups" that is a categorical variable to group my values together. What formula can I use in a third column to get the mean of "myValues" group by group.

E.g.

myValues	myGroups	myMeans
1	A	2
2	A	2
3	A	2
4	B	5
5	B	5
6	B	5
7	C	7.5
8	C	7.5
9	D	9

Thank you

Re: Formula for grouped statistics

The "ColStat" fumctions is mainly useful in column formulas and cannot return non-repeated values as far as I know.

But Summarize() can. Example:

Summatize(g = group(:mygrps), m = Mean(my values))

Where m is matrix of the mean at each level of g, a list of group names (as strings). Excluded rows are not included.

Re: Formula for grouped statistics

if you give an example perhaps i can understand better what you are looking for.

Re: Formula for grouped statistics

Hi Ron,

yes, that would produce the same result. so i was asking if JMP can have non repeated data functionality in col mean( my values, mygrps).

Thanks

ram

Re: Formula for grouped statistics

hi ram.asra@gmail.com,

for unique values try using Table Summary.

by the way, summarizing the table and then updating the original table can produce the same result as guilhem was asking for just without the formula.

ron

Re: Formula for grouped statistics

Can it give only one unique value, instead of repeated data points?

Re: Formula for grouped statistics

Thanks!

Re: Formula for grouped statistics

Use the following formula: Col Mean(:myValues, :myGroups)