Using function handles and indexing on cells, Error: Undefined function 'eq' for input arguments of type 'cell'.

Hello! Thank you for reading my question!

I have a dataset that is the following:

DO_data_and_quality =

    [226051x10 double]    [226051x1 double]    {226051x1 cell}    {226051x1 cell}

First: station name, second:datenum, third: dissolve oxygen measurements (has NaN because of missing fields due to spacing or missed measurement), fourth: error code in the form of the following:

%<<0>> {a number -5:1:5 enclosed by <<>>}
%<<0>>[ERR] {the above and a bracketed three letter error code}
%NaN  {for empty cells, it came from an excel file}

Primary objective: Remove the a row that has NaN, for all of the cell objects. Secondary objective: To group the error codes into their own cells, with the corresponding row of the other cells

*Code for primary objective:

Flag_DO=NDBCdata.F_DO_mgl;
rawDO=NDBCdata.DO_mgl;
fh1 = @(x)isnan(x(:));
whereNaN_Flag_DO=cellfun(fh1, Flag_DO,'UniformOutput',false); 
whereNaN_DO=cellfun(fh1, raw_DO,'UniformOutput',false); %creates a vector where 1 means that there is a NaN, 0's mean that there isn't one

 % if =1 has a NaN then =0 there is no NaN, then any number besides 0 has a NaN, remove the row
  raw_DO=NDBCdata.DO_mgl;
  fh2=@(x)(x(whereNaN_DO ==1 | whereNaN_Flag_DO ==1));
  DO_data_and_quality={ NDBCdata.ID_num NDBCdata.DateNUM raw_DO Flag_DO};

DO_data_and_quality(cellfun(fh2, DO_data_and_quality)) = [];

I am not using indexing correctly, may I have a few pointers?

Even more stangely when I click on the problem line in the error msg I get: Undefined function or variable 'whereNaN_DO'.

I did a whos whereNaN_DO and I get an output! Name Size Bytes Class Attributes

whereNaN_DO      226051x1             25543763  cell

*For secondary objective

I was going use: codeFlag_DO=cellfun(@double,Flag_DO, 'UniformOutput', false); and use a similar indexing method where I would select the second term or the last few terms if there is a 3 letter error code.

Thanks!

[Cathleen Turner]

To clarify what I need to do...

if I were to do this in a for loop I would add:

mat1=cell2mat(whereNaN_Flag_DO);
mat2=cell2mat(whereNaN_DO);

itrs=size(Flag_DO);
itrs=max(itrs);
itrs=size(Flag_DO);
itrs=max(itrs);

for gg=1:itrs
    if mat1(gg) ==0 | mat2(gg) ==0
        Store{gg}=Flag_DO{gg}
    end
end

I don't like this method because 1) it despairs data by giving each string its own cell, and creates many many columns 2) I want to do it using indexing. If 2 isn't feasible how can fix the problem with 1)?

[Cathleen Turner]

edit 2:

to give you an idea of what the error dataset is like

 Flag_DO{1:2}
ans =

<0>

ans =

<0>[GFT]

the problem with the forloop is that it would separate the second answer into 2 columns and I end up having many many columns

Answer 1

[Walter Roberson]

My guess is that you want

fh1 = @(x) any(isnan(x(:)));

and that for whereNaN_DO you want to use uniform true

[Cathleen Turner]

Thank you for your response. However I get this error and I am not sure how to respond:

Undefined function 'eq' for input arguments of type 'cell'.

Error in @(x)(any(whereNaN_DO==1|whereNaN_Flag_DO==1))

[Tom]

x isn't used anywhere in that anonymous function?

[Cathleen Turner]

Oh I thought that you had to write a function handle with x by default then use it for cell function. Was I wrong?

[Tom]

It works similar to a standard function - if you don't need any input arguments (which is a bit unusual), then you'd write it as @() stuff.... ; if you need to arguments (say for a uicontrol callback), you could write @(hObject,eventData) stuff...