Can one use retime function to calculate mean and sum of the same variable?

Suppose I have a timetable

                            Visitors
January 1                  10
January 3                  20
February 10                30
February 12                10

Then I want to create a table

                                                Mean of Visitors     Sum of Visitors
January                                           15                           30
February                                          20                          40

Can one do this with retime function? if not, how?

[Torsten]

Shouldn't Mean of Visitors be 30/31 since January has 31 days ?

[alpedhuez]

Since there are two observations for each month. Want to calculate an average per observation.

[Image Analyst]

What about the days of the month not listed there? Do we assume the venue was closed on missing days or that the number of visitors those days was 0? Have you seen these functions: splitapply, grpstats, and groupsummary?

[dpb]

Yes, it can be done with retime to compute multiple outputs (there are examples of doing so in the doc) but for this look at groupsummary

[alpedhuez]

Probably better to use grpstats function.

[dpb]

grpstats is in Statistics TB; groupsummary is (since R2018a) in base product. Have overlapping functionality on builtin statistics available with some differences but the basics are in both. groupsummary has additional functionality with the 'groupbins' binning scheme that mimics retime type functionality that grpstats doesn't have builtin at all.

Another case where the TB folks and the main product folks step on each others' toes and there's way too much duplication/overlap/redunancy/bloat/...a TB lets some new feature out into the wild; it's popular and picked up in the base product but then there's never a way to clean up the debris field left over...

[alpedhuez]

I would think of these data types as another example that maybe getting too complicated.

[dpb]

???? You mean functions, @alpedhuez?

[alpedhuez]

As another example that may getting a bit too complicated.

Answer 1

[Gagan Agarwal]

Hi alpedhuez,

Yes, it is possible to calculate the mean and sum of the same variable using the 'retime' function.

You can refer to the following code snippet as an example to perform the calculation on your table:

% The sample data is stored in a table 'T' 
T = table(["January 1"; "January 3"; "February 10"; "February 12"], [10; 20; 30; 10]); 
T.Properties.VariableNames = {'Date', 'Value'}; 
T.Date = datetime(T.Date, 'InputFormat', 'MMMM d') 
% T is converted to timetable using 'table2timetable' function. 
TT = table2timetable(T); 
% 'retime' function is then applied on 'TT' to calculate the mean 
T1 = retime(TT, 'monthly', 'mean'); 
% The 'retime' function is then applied on 'TT' to calculate the sum 
T2 = retime(TT, 'monthly', 'sum'); 
% 'T1' and 'T2' are combined into a single table called 'ans' which displays the sum and mean values 
ans = join(T1,T2) 

For more detailed information about the 'retime' function, you can refer to the following documentation:

https://www.mathworks.com/help/matlab/ref/timetable.retime.html

I hope this helps!