Retime mean data not aligning

Hi there,

I'm using retime on a data set with a one second interval, to obtain averages for days and months (and many other intervals). One of the variables, Cool, in the table is a cooling % calculated from two other table variables, C1 and C2;

T1.Cool = 100*((T1.C1 - T1.C2) ./ T1.C1);

However when I use retime to change the interval of the data e.g.

TDay = retime(T1,"daily","mean");

The values of TDay.Cool is no longer the cooling difference between the new TDay.C1 and C2 variables. I.e. If Tday.C1 =100 and Tday.C2 = 90, Tday.Cooling shows as 9% rather than the actual cooling which is 10%.

What is happening in retime to get such behaviour and how can I correct it?

[Peter Perkins]

Oliver, I think you are going to need to post a small example of your data. if Tday.C1=100, Tday.C2=90, and you set Tday.Cooling = 100*((Tday.C1 - Tday.C2) ./ Tday.C1), 9% seems impossible to end up with.

[Oliver Higbee]

@Peter Perkins, Cooling is calculated on the data with a one second interval (and is correct for that data). Then using the retime to adjust to a daily mean I get the above table. The only difference between Cooling and Correct Cooling is that Correct Cooling is calculated after retime.

Given that I'm seeing this behaviour, I don't have confidence that retime is giving the correct means for C1 and C2 either. Any ideas what is happening?

[Walter Roberson]

I think we will need a data table to test with.

[Oliver Higbee]

Sure, apologies for not starting with this!

Attached are the input and output files (Had to retime to 1minute to get below attachment file size). With just the following code used to get the output;

TDay = retime(T1,'daily','mean'); %Take daily averages
TDay.("Correct Cooling (%)") = 100*((TDay.C1 - TDay.C2) ./ TDay.C1);

[Peter Perkins]

Oliver, are you computing

mean(100*((T1.C1 - T1.C2) ./ T1.C1))

and expecting that to be the same as

100*((mean(TDay.C1) - mean(TDay.C2)) ./ mean(TDay.C1))

? Those are not the same thing.