how to user grpstats function in this case

I have a table like (the real table is more complicated)

date month person spending
1/1/2020 January-2020 John 30
1/1/2020 January-2020 Mike 15
1/2/2020 January-2020 John 20
1/3/2020 January-2020 Kate 30

then I want to create a summary table such as

person 'count of John's spending data' 'sum of John's spending' 'average of John's spending per observation'
January-2020
February-2020

Please advise.

Answer 1

[dpb]

>> grpstats(thuey,{'person','month'},{'mean','@sum'},'DataVars',{'spending'})
ans =
  3×5 table
                     person     month      GroupCount    mean_spending    Fun2_spending
                     ______    ________    __________    _____________    _____________
    John_Jan-2020     John     Jan-2020        2              25               50      
    Kate_Jan-2020     Kate     Jan-2020        1              30               30      
    Mike_Jan-2020     Mike     Jan-2020        1              15               15      
>> 

[alpedhuez]

Not exactly asked in the question.

[dpb]

Well, you'll have to do some work yourself, then.

grpstats does what grpstats does...

[alpedhuez]

create a table for John. rename the column name. do it also for Kate and Mike. Then merge all three tables?

[dpb]

You've got all the results in the above table; just rearrange the pieces and output in the format wanted.

It's essentially your desired table with the added group label column of both person and month.

[dpb]

>> tapled=table(tgrps.person,tgrps.month,tgrps.GroupCount, ...
                tgrps.mean_spending,tgrps.Fun2_spending, ...
                'VariableNames',{'Individual','Time Period','Number Outlays', ...
                'Total Outlay','Average Outlay'})
tapled =
  3×5 table
    Individual    Time Period    Number Outlays    Total Outlay    Average Outlay
    __________    ___________    ______________    ____________    ______________
       John        Jan-2020            2                25               50      
       Kate        Jan-2020            1                30               30      
       Mike        Jan-2020            1                15               15      
>> 

To write variable text would require building those strings via sprintf or compose or the like...it's doable but probably not a fun thing to code.

NB: To organize by individual instead of month you'd want to reorder the grouping variables or else will have to use varfun or the like to regroup in that order instead.

NB2: Wasn't able to find a way to turn off the 'RowNames' property of the original table once they were written by grpstats; hence the table construction above. Hadn't ever noticed that "feature" before...

[dpb]

>> fmt="Number %s's outlays, total of %s's outlays, average of %s's outlays";
>> compose(fmt,repmat(tapled.Individual,1,3))
ans = 
  3×1 string array
    "Number John's outlays, total of John's outlays, average of John's outlays"
    "Number Kate's outlays, total of Kate's outlays, average of Kate's outlays"
    "Number Mike's outlays, total of Mike's outlays, average of Mike's outlays"
>> 

let's you build the sample string; I don't know just what you really envision as the output you're looking for. The table above seems pretty good to me; I don't see the purpose in the repetition.