How to assign a double value to a cell array?

My goal is to replace any missing values in my cell array T1 with -999, if the column is composed of floating numbers, and replace any missing values with an empty string like ' ', if the column is composed of text strings?

Below is my code. Unfortunately, I got an error saying: "Conversion to cell from double is not possible."

% Replace the remaining missing values with -999: 
missingIndices2 = ismissing(string(T1));
T1(missingIndices2) = -999;      
%  convert the cell into a table:
Table = cell2table(T1);

Answer 1

[Voss]

Put curly braces {} around the -999, which makes the right-hand side of the assignment a scalar cell array:

T1(missingIndices2) = {-999};

[Leon]

Thanks for the suggestion. I just tried it. Unfortunately, my table column turns into something like the below:

{[ -999]}

{[ -999]}

{[ -999]}

{[35.9963]}

{[35.8279]}

When it should be

-999

-999

-999

35.9963

35.8279

[Voss]

Please upload a mat file with the cell array T1.

[Stephen23]

"Unfortunately, my table column turns into something like the below: "

It works here:

C = {-999; -999; -999; 35.9963; 35.8279};
T = cell2table(C)

T = 5x1 table
      C   
    ______

      -999
      -999
      -999
    35.996
    35.828

[Leon]

Please see attacned for a mat file containing the cell array T1. Many thanks for your help.

[Stephen23]

C = load('test.mat').T1

C = 1498x2 cell array
    {[35.7030]}    {[35.7010]}
    {[35.5790]}    {[35.5780]}
    {[35.5790]}    {[35.5790]}
    {[34.7100]}    {[34.7090]}
    {[34.7180]}    {[34.7190]}
    {[34.7170]}    {[34.7130]}
    {[34.9420]}    {[34.9410]}
    {[34.3970]}    {[34.3820]}
    {[34.3970]}    {[34.3820]}
    {[34.2390]}    {[34.2390]}
    {[34.2390]}    {[34.2390]}
    {[34.0750]}    {[34.0710]}
    {[34.0760]}    {[34.0670]}
    {[33.4280]}    {[33.4240]}
    {[33.4280]}    {[33.4230]}
    {[36.0710]}    {[36.0740]}

C(~cellfun(@isscalar,C)) = {-999};
T = cell2table(C)

T = 1498x2 table
      C1        C2  
    ______    ______

    35.703    35.701
    35.579    35.578
    35.579    35.579
     34.71    34.709
    34.718    34.719
    34.717    34.713
    34.942    34.941
    34.397    34.382
    34.397    34.382
    34.239    34.239
    34.239    34.239
    34.075    34.071
    34.076    34.067
    33.428    33.424
    33.428    33.423
    36.071    36.074

This is probably better handled when the data is imported.

See also: https://www.mathworks.com/matlabcentral/fileexchange/19730-cell2float

[Leon]

Many thanks.

There is one thing though. The code uses -999, why are the output NaNs instead? Additionally, the T1 I shared was a simplified one. In reality I would have many columns that are text strings as well. Is there a way I could replace missing values in float data columns with -999, but replace missing values in string columns with an empty string like ' '? Thanks.

C(~cellfun(@isscalar,C)) = {-999};

[Voss]

"why are the output NaNs"

Because your cell array has some cells that contain scalar numerics, some cells that contain <missing>s, and some cells that contain (empty) character vectors.

If you use ismissing(string(C)) as the condition to replace with {-999}, then you replace the <missing>s but not the empty character vectors because string('') is "" and "" is not <missing>.

ismissing(string(''))

ans = logical
   0

If you use ~cellfun(@isscalar,C) as the condition, then you replace the empty character vectors but not the <missing>s because <missing> is a scalar

isscalar(missing)

ans = logical
   1

To replace the contents of any cell that contains <missing> or a non-scalar array, which seems like what you are trying to do [EDIT: not really, thanks for clarifying], you can use some condition like cellfun(@(x)~isscalar(x) || ismissing(x),C)

C = load('test.mat').T1;

rows = [99:108, 576:585];

C(rows,:)

ans = 20x2 cell array
    {[33.8186]}    {[<missing>]}
    {[33.8898]}    {[  33.8903]}
    {[33.9370]}    {[  33.9366]}
    {[33.9753]}    {[  33.9761]}
    {[33.9749]}    {[<missing>]}
    {[33.9249]}    {[  33.9261]}
    {[33.9613]}    {[  33.9631]}
    {[34.1909]}    {[  34.1920]}
    {[34.2087]}    {[<missing>]}
    {[34.2238]}    {[<missing>]}
    {[36.3150]}    {[  36.3090]}
    {[36.3190]}    {[  36.3130]}
    {[36.3410]}    {[  36.3340]}
    {[36.2880]}    {[  36.2920]}
    {[36.1500]}    {0x0 char   }
    {[36.6150]}    {[  36.5980]}

C(cellfun(@(x)~isscalar(x) || ismissing(x),C)) = {-999};
T = cell2table(C);

C(rows,:)

ans = 20x2 cell array
    {[33.8186]}    {[   -999]}
    {[33.8898]}    {[33.8903]}
    {[33.9370]}    {[33.9366]}
    {[33.9753]}    {[33.9761]}
    {[33.9749]}    {[   -999]}
    {[33.9249]}    {[33.9261]}
    {[33.9613]}    {[33.9631]}
    {[34.1909]}    {[34.1920]}
    {[34.2087]}    {[   -999]}
    {[34.2238]}    {[   -999]}
    {[36.3150]}    {[36.3090]}
    {[36.3190]}    {[36.3130]}
    {[36.3410]}    {[36.3340]}
    {[36.2880]}    {[36.2920]}
    {[36.1500]}    {[   -999]}
    {[36.6150]}    {[36.5980]}

T(rows,:)

ans = 20x2 table
      C1        C2  
    ______    ______

    33.819      -999
     33.89     33.89
    33.937    33.937
    33.975    33.976
    33.975      -999
    33.925    33.926
    33.961    33.963
    34.191    34.192
    34.209      -999
    34.224      -999
    36.315    36.309
    36.319    36.313
    36.341    36.334
    36.288    36.292
     36.15      -999
    36.615    36.598

[Leon]

Working now. Thank you so much for the tremendous help!