xlsread doesn't recognize file has changed

I have a .xlsx file that I read with xlsread. However, if I change the file and call it with xlsread again, it returns the old values. I've tried clear all, clear('myfile.xlsx'), pack, fclose all, but nothing seems to make the xlsread read the newly modified file, except for restarting matlab. How can I force it to read in the new file? And yes, I have saved the new file.(using 2015asp1)

[dpb]

>> B=xlsread('B.xls')
B =
 36.3027   51.2517
 36.2643   51.2544
 36.2613   51.2665
>> B1=B(1);  % save old value to later restore
>> B(1)=pi
B =
  3.1416   51.2517
 36.2643   51.2544
 36.2613   51.2665
>> xlswrite('B.xls',B)  % write the modified back
>> clear B              % ensure isn't cached value in memory...
>> B=xlsread('B.xls')   % and see what happens...
B =
  3.1416   51.2517
 36.2643   51.2544
 36.2613   51.2665
>> B(1)=B1;
>> xlswrite('B.xls',B)
>> xlsread('B.xls')
ans =
 36.3027   51.2517
 36.2643   51.2544
 36.2613   51.2665
>>

If can't reproduce the above test or something very similar, submit the test results to TMW official support at www.mathworks.com

Above is from R2012b here but I'd surely expect same result from later revisions, too...

ADDENDUM

I also went and manually changed the spreadsheet and Save'd if from within Excel and read it after each change -- first for the first cell modified from initial value to '=pid()' and then restored the initial value. As expected, the results were --

>>  % Modify in Excel and read...
>> xlsread('B.xls')
ans =
  3.1416   51.2517
 36.2643   51.2544
 36.2613   51.2665
>> % Put numeric value back, save and re-read...
>> xlsread('B.xls')
ans =
 36.3027   51.2517
 36.2643   51.2544
 36.2613   51.2665
>>

Seems fine here.

[John Petersen]

It looks like it may have been a problem with excel, because after I saved the file then closed it, I reopened it and it had the old data again -- not the recently "saved" data. I edited again and now it seems to be behaving. properly.

[dpb]

Ah so...I've had some "issues" with Excel, too, so wouldn't surprise me. Perhaps there was another session that hadn't closed properly and was holding the file handle or the like...