Can't get all the rows in my table that have any missing values and remove those rows

Hi there,

I am trying to remove my table (of a combination of double, cell arrays and duration arrays) of any rows which have missing values. For example, a subset of the table is:

''  ''  '441'  '0'
''  ''  '5'  '0'
'No'  ''  '2889'  '0'
'No'  'Clicks'  '2894'  '0'
''  ''  ''  '0'
''  ''  '5'  '0'
'No'  'ROD'  '3591'  '1'
''  ''  ''  '0'

I tried this link.

My code is:

badRows = ismissing(finalnbs);
T = finalnbs(~badRows, :);

However, this results in an error:

Row index exceeds table dimensions.

What am I doing wrong, and what code do I need to achieve my need?

[the cyclist]

Could you upload one of the following?

• A *.mat file with the actual table
• Code that reproduces the table, or at least a subset of it that recreates the problem

[Image Analyst]

Yes, it's always a good idea for posters to give complete code that we can copy and paste, not like what you did. So, my question is how can you have 5 or more missing values in a row when the rows of your table have only 4 columns?

[Guillaume]

Also, don't put questions in the title, particularly if it is a different question than in the body.

Both the cyclist and I answer your question in the body (remove any row which has missing value). The answer to remove any row that has more than x missing values is completely different

[Dhruv Ghulati]

Apologies, you are right, here the question in the title doesn't match the question in the body given. For the sake of other readers I am amending the title.

Answer 1

[the cyclist]

This should solve it for you:

A = {
''  ''  '441'  '0';
''  ''  '5'  '0';
'No'  ''  '2889'  '0';
'No'  'Clicks'  '2894'  '0';
''  ''  ''  '0';
''  ''  '5'  '0';
'No'  'ROD'  '3591'  '1';
''  ''  ''  '0'};

T = cell2table(A)

badEntry = ismissing(T);

badRow = any(badEntry,2)

T = T(~badRow, :);

Answer 2

[Guillaume]

badRows is an m*n logical array, where m is the number of rows and n the number of columns of your table. So it has m*n elements. You're then using that as a row index into your table, so of course, matlab can't fit m*n values into m values.

To fix this, use the any (or all) function:

T = finalnbs(~any(badrows, 2), :);
%or
T = finalnbs(all(~badrows, 2), :);

P.S.: this answer your question on how to remove any row which has missing value. If you want to remove rows have more than x missing values, then you use sum instead:

T = finalnbs(sum(badrows, 2) <= x, :);