matlab how use format bank in cell arrays

I am sorry if I repeated any of the questions, but I did not find an answer to my problem anywhere.

I would like the data displayed in my cell array to be in a decimal place.

I used the bank format but it doesn't work.

I tried several solutions with rounding, tahat I found here, but I got the error that the solution is only possible for a 2D table.

I also tried to assign a format to individual rows in the arrays - it doesn't work

I have an cell array of data retrieved from outside (% f)

I need to paste some data to Excel and when I copy it I get the result:

0.763361344537815 6.71428571428571 -4.43000000000000 -0.0363619047619048 0.170616560896359 10 100 2.79400000000000 1.43163922691482

I need many files and formatting in Excel will take a long time, that's why I wanted to change the format directly in matlab.

Answer 1

[Guillaume]

There's several things here:

Firstly, format only affects the way matlab displays numbers at the command prompt. It does not affect what is stored, and it does not affect how they're displayed in the variable editor. Since it does not affect what is stored, a corrolary is that it does not affect the way numbers are exported to excel, text files, or any other file format.

If you set format bank and display your cell array at the command line (with e.g celldisp) then numbers will indeed show with 2 decimals. If that's not the case for you, then you've found a bug. But again, that doesn't affect the numbers themselves, matlab still store them to full precision (~16 decimal digits)

If you want to change the precision with which the numbers are stored, you do need to use round. If your data is stored in a cell array, then round needs to be applied to each cell of the cell array, not to the cell array itself. It's not very clear what you have exactly, something like:

roundedcellarray = cellfun(@(m) round(m, 2), yourcellarray, 'UniformOutput', false);

may work.

You should not be pasting data between matlab and excel (or vice-versa), there are much more reliable ways of transfering data between the two, such as writetable, readtable, writematrix, readmatrix, etc. Again, it's not very clear what you are doing.

[Caroline]

I tried this method and rounded, but it adds zero. So I don't have 2 decimal digits.

I copy to Excel because I do a summary in Excel and I only need some data from different cell arrays, made on different days. Some measurements are bad and I reject them. It's faster, but I have a problem with the format and I thought it could be improved.

[Guillaume]

Please describe in details the exact workflow you're using as I'm a bit confused.

Again, there's a difference between storage and display. round changes how digits of precision are stored (roughly) but doesn't affect the display. You can round to 2 digits but still display with 5 digits (3 of which are 0) or just one digit.

If you're doing copy/paste from the command window to excel, then format bank is indeed what you want. If you're doing something else, then please explain.

Also, note that whatever you do in matlab, doesn't affect the cell formatting in excel. That's controlled by excel itself.

[Stephen23]

"If you want to change the precision with which the numbers are stored..."

Then you need to change numeric class, rounding makes no difference to the storage precision.

[Guillaume]

@stephen, you didn't include my (roughly) in your quote which was meant as some big hand-waving to avoid talking about floating point representation. For the purpose of changing 0.763361344537815 to (again hand-waving away actual FP representation) 0.76 rounding would work.

But it's really not clear what caroline is after.

Answer 2

[Caroline]

I choose the data I need from previously saved workspace cell arrays and transfer to Excel.

I thought I could do matlab formatting, but it would be better in Excel.

Thank you for the guidance :)