Facing this weird comparison error of vector
Ältere Kommentare anzeigen
I have some data that needs to be read from a file.
contents of file are as follows:
point1
0,0,2.2
point2
1,0,2.2
so i have used the fread command to read contents and used the Newline command to store each line as a member of cell(4x1 in this case).
I used the following command to convert the point data in form of string into vector
A = str2num(strtrim(Lines{2}))
D = str2num(strtrim(Lines{4}))
Outputs: (Double)
A = [0,0,2.2000]
D = [1,0,2.2000]
Here comes the error when i tried to compare the 3rd elements of A & D and to my surprise the following comparison expression false:
A(3) == D(3)
ans = 0 (logical)
Can anybody say what is wrong in this that causing this type of result ?
Please share your thoughts
7 Kommentare
Mathieu NOE
am 26 Okt. 2020
hi
I remember some info about str2num :
help str2num CAUTION: str2num uses EVAL to convert the input argument, so side
effects can occur if S contains calls to functions. Use
STR2DOUBLE to avoid such side effects or when S contains a single
number. Also spaces can be significant. For instance, str2num('1+2i') and
str2num('1 + 2i') produce x = 1+2i while str2num('1 +2i') produces
x = [1 2i]. These problems are also avoided when you use STR2DOUBLE.:
so please try with STR2DOUBLE
Stephen23
am 26 Okt. 2020
"Can anybody say what is wrong in this that causing this type of result ?"
Most likely some floating point error, i.e. those two values are not exactly the same.
Mahesh Babu Dhanekula
am 27 Okt. 2020
Stephen23
am 27 Okt. 2020
@Mahesh Babu Dhanekula: please upload the data file by clicking the paperclip button, and show us the exact code that you use to import and compare the data.
Mahesh Babu Dhanekula
am 27 Okt. 2020
Bearbeitet: Mahesh Babu Dhanekula
am 27 Okt. 2020
"I forgot to add a point about i am modifying one point to get the vertices of geometry"
Ah, secret operations that accumulate floating point error :)
"Here after i am facing this comparison error as you said it might be attributed to floating point error."
Yes.
Use this to look at the exact values, you will see that they are not the same:
Tip: do not test floating point numbers for exact equivalence. Always compare the absolute difference against a tolerance:
abs(X-Y)<tol
Mahesh Babu Dhanekula
am 27 Okt. 2020
Antworten (0)
Kategorien
Mehr zu Data Type Conversion finden Sie in Hilfe-Center und File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!