When I convert my data which is available in structure format to table it accuracy decreases

When I convert my data which is available in structure format to table which I requre for further plotting the values like 0.7 or 0.5 changes to 1 and 0.

I am use the command

struct2table(name);

for example

from

 0.505829210007949
0.549192911465458
0.539702546882381
0.529415778849225
31.7804971641414
0.459445413396004
1.18079450595063
1.21475111433387

to

0
0
0
0
1

[Ive J]

Please provide your script along with the result you get. I cannot reproduce your output solely based on your question.

myStruct.vals  = [0.505829210007949
0.549192911465458
0.539702546882381
0.529415778849225
31.7804971641414
0.459445413396004
1.18079450595063
1.21475111433387];

struct2table(myStruct)

ans =

  8×1 table

       vals   
    _______

    0.50583
    0.54919
     0.5397
    0.52942
      31.78
    0.45945
     1.1808
     1.2148

[dpb]

No, you did something else either in the display formatting or converting to a logical array or somesuch. MATLAB will retain full precision of the data in the conversion internally; there may be some slight loss of precision if you were to write to a file with high-level routines but it would still have 5-6 decimal digits of precision at worst.

You'll have to post a complete working example that reproduces the problem for folks to be able to diagnose just exactly what you did to get the symptom, but it is operator error, not MATLAB bug.

Answer 1

[Walter Roberson]

Your table contains an integer datatype. When you combine numeric datatypes in an array, the result is converted to the most restrictive datatype

val = [int8(-3) int16(555) pi]

val = 1×3

-3 127 3

●

class(val)

ans = 'int8'

What shows up for

unique( varfun(@class, T, 'output', 'cell') )