Creating subset of table

When I am trying to create a subset of my table like so:

table_1(table_1.col_1>=0,:)

I am getting this error:

Operator '>=' is not supported for operands of type 'cell'

col_1 contains numbers, but they might not be recognized as numerical var type. Is there an elegant solution to this.

I have been trying strcmp instead, but that does not work on numerical comparisons.

I also tried to convert the column to an array using cell2mat but then I ned up getting an error saying:

Row index exceeds table dimesions.

[Stephen23]

"col_1 contains numbers, but they might not be recognized as numerical var type. Is there an elegant solution to this."

The first thing to try is STR2DOUBLE.

If that does not work then upload your data in a .mat file by clicking the paperclip button.

Answer 1

[Scott MacKenzie]

Since column 1 contains numbers in cells, something like this is needed:

T = array2table([{1, -2, 3}' {'a' 'b' 'c'}']) % test data

T = 3×2 table
     Var1     Var2 
    ______    _____

    {[ 1]}    {'a'}
    {[-2]}    {'b'}
    {[ 3]}    {'c'}

T.Var1 = cell2mat(T{:,1})

T = 3×2 table
    Var1    Var2 
    ____    _____

      1     {'a'}
     -2     {'b'}
      3     {'c'}

T(T.Var1 >= 0,:)

ans = 2×2 table
    Var1    Var2 
    ____    _____

     1      {'a'}
     3      {'c'}

[Metin Akyol]

Thank you for that. Didn't quite work for me though, still getting the same error. That transformation that you suggested, still leaves my column unchanged, i.e. it is still a cell array.

[Scott MacKenzie]

Ok, I just did a little edit to the answer. I'm assuming column 1 contains numbers only in cells.

[Metin Akyol]

I am still getting the same error saying that Row index exceeds table dimensions when I use cell2mat. In fact, using cell2mat converts the column to a struct with fields on my end.

[Metin Akyol]

Your initial table T is also different than my starting table, in the sense that your table has T.Var1 being a 3x1 cell, my column is a cell array.

[Scott MacKenzie]

Moving forward, it would be best if you post your data and the code that generates the error. Of course, you can also just post a subset of your data along with code demonstrating the error.