Replacing Empty Cells by NaN

Hi Matlab Community,

I would appreciate some help. I am importing an excel .csv file into a table. This file has multiple cells as empty cells. However, once imported as table in Matlab, some empty cells are filled with NaN while others show {0×0 char}. (Refer below)

I want to replace all empty cells in the table with NaN value, because using table2array function with {0×0 char} gives an error, which is as follows:

Error using table2array (line 37)

Unable to concatenate the table variables 'Var1' and 'Var16', because their types are double and cell.

Error in Berea_NTW1 (line 42)

node1 = table2array(node1Table) ;

Thanks a lot !!

[Ive J]

Woud you mind to share your data file? use paper clip to upload.

Answer 1

[Scott MacKenzie]

You can use fillmissing and specify the fill value for each column. Obviously, the fill values depend on the data type in the column. Here's an example:

T = table({'a' '' 'f' 'z'}', [1 2 nan 4]', [9 nan nan 5]', {'z' '' '' ''}')

T = 4×4 table
       Var1       Var2    Var3       Var4   
    __________    ____    ____    __________

    {'a'     }      1       9     {'z'     }
    {0×0 char}      2     NaN     {0×0 char}
    {'f'     }    NaN     NaN     {0×0 char}
    {'z'     }      4       5     {0×0 char}

T = fillmissing(T, 'constant', {'Oops!', 0, -1, 'Unknown'})

T = 4×4 table
      Var1       Var2    Var3       Var4    
    _________    ____    ____    ___________

    {'a'    }     1        9     {'z'      }
    {'Oops!'}     2       -1     {'Unknown'}
    {'f'    }     0       -1     {'Unknown'}
    {'z'    }     4        5     {'Unknown'}

[Muhammad Farooq Zia]

Thank you for your reply. I tried putting NaN as the constant value in fillmissing function, but still its not working. This is the code line where I try this:

node1Table = fillmissing(node1Table, 'constant', NaN) ;

can you please tell me how I can resolve this?

[Scott MacKenzie]

You don't insert NaN in fillmissing. You insert the value you want to substitute for NaN, such as 0. Just look at my example code.

Also, since the columns in your table have different data types, you need to specify the replacement values according to the type of data in each column; i.e., you cannot use NaN as the replacement value for a column that contains char data. Again, see my example code.

Answer 2

[Simon Chan]

T.Var1=NaN(size(T.Var1,1),1);                           % Replace T.Var1 by NaN
T.Var16=NaN(size(T.Var16,1),1);                         % Replace T.Var16 by NaN