Replace NaN with Blanks

175 views (last 30 days)
Abdullah Azzam
Abdullah Azzam on 7 Feb 2020
Commented: David on 5 Jun 2024
Hi guys I have a double matrix as shown in the attached picture that I am trying to replace all NaN values with Empty spaces so that the over all matrix dimension will be maintained. However, when I looked through the function and suggested solutions non of them is actully working what possible code can be used.
Matrix.PNG
Thanks in advance
  1 Comment
David
David on 5 Jun 2024
This is not a "fix it in Matlab" answer, but I've been generating tables in Matlab to use in Word for example. When I build the table matrix, I insert NaN's in every cell that I want blank. Then in Word, I use "find and replace" to replace all those with something else; i.e., '---'. This idea should work in Excel too, although I've not tested it there.

Sign in to comment.

Accepted Answer

Walter Roberson
Walter Roberson on 7 Feb 2020
Edited: Walter Roberson on 7 Feb 2020
It is not possible to have a numeric array that has blanks in it.
You can use
fillmissing(V1, 'constant', 0)
to replace the NaN entries with 0... or you could do
V1(isnan(V1)) = 0;
for the same effect.
If you need a blank then you will need to convert to cell array, such as
V1c = num2cell(V1);
mask = cellfun(@(C) all(isnan(C)), V1c);
V1c(mask) = {' '};
If you do this, then the variable browser will show ' ' (including the quotes) in those locations but the quote marks are not part of the content. disp(V1c) would show ugly things like
{[ 1]} {[ 1]} {[ 1]} {[2]} {[3]}
{' '} {[ 3]} {[ 2]} {[1]} {[3]}
{[ 2]} {' '} {[ 3]} {[3]} {[2]}
{[ 3]} {[ 2]} {[ 3]} {[3]} {[2]}
{[ 3]} {[ 3]} {' '} {[3]} {[1]}
  4 Comments
Abdullah Azzam
Abdullah Azzam on 8 Feb 2020
Ok. Then how may I import an excel sheet that contain both numaric and string value into the matlab. Note if it isn't possible I don't mind if the numric were saved as strings
Walter Roberson
Walter Roberson on 8 Feb 2020
readtable() is what is recommended these days.
Or you could xlsread() . The first input will contain only numeric parts, the second output will contain text without numbers, and the third will contain both. But readtable() is typically better.
In some cases it might be more suitable to use readcell() for R2019b and later.

Sign in to comment.

More Answers (2)

dpb
dpb on 7 Feb 2020
Can't be done--a double array has to be regular in both dimension; can't have holes.
Only possibility is to convert to a cell array.

Opeyemi Kehinde
Opeyemi Kehinde on 5 Dec 2021
fillmissing(V1, 'constant', [])
will this work to make NaN an empty entry
  2 Comments
Walter Roberson
Walter Roberson on 5 Dec 2021
Let us test:
V1 = [1 2 nan 3; 4 nan 5 6]
V1 = 2×4
1 2 NaN 3 4 NaN 5 6
fillmissing(V1, 'constant', [])
Error using fillmissing/checkConstantsSize (line 740)
Fill constant must be a scalar or a 4-element vector.

Error in fillmissing/parseInputs (line 680)
intConstOrWindowSize = checkConstantsSize(A,AisTable,false,intConstOrWindowSize,dim,dataVars,'');

Error in fillmissing (line 158)
[A,AisTable,intM,intConstOrWinSize,extM,x,dim,dataVars,ma,maxgap] = parseInputs(A,fillMethod,varargin{:});
So... NO.

Sign in to comment.

Categories

Find more on Data Preprocessing in Help Center and File Exchange

Products


Release

R2017a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!