CODEWITHSUNDEEP
arraysmatlabmatrixreplacenan
Markus
Markus

Reputation: 33

Replace all NaN in specific row with same value - Matlab

I have an array (data2winner) that looks like this:

Columns 1 through 6

1.0894    1.1923    1.0238    1.0625    1.0222    0.9701
1.0000    1.2419    1.2093    0.9412    0.9783    1.0923
1.0379       NaN    1.0000    1.0417    1.0667    1.0000

Columns 7 through 12

0.9661    0.9568    1.0400    0.8519    0.8675    0.9821
0.7895    1.0000       NaN    1.0000    1.1250    0.8182
1.0667    0.7714       NaN    1.0000    0.9383    1.1111

I want to replace all NaN in the same row (e.g. row 3) with the same value, lets say 0.5, such that the output looks like this:

Columns 1 through 6

1.0894    1.1923    1.0238    1.0625    1.0222    0.9701
1.0000    1.2419    1.2093    0.9412    0.9783    1.0923
1.0379       **0.5**    1.0000    1.0417    1.0667    1.0000

Columns 7 through 12

0.9661    0.9568    1.0400    0.8519    0.8675    0.9821
0.7895    1.0000       NaN    1.0000    1.1250    0.8182
1.0667    0.7714       **0.5**    1.0000    0.9383    1.1111

I know I can replace all NaN in my matrix by 0.5 using this:

data2winner(isnan(data2winner)) = 0.5

So i thought using

data2winner(isnan(data2winner(3,:))) = 0.5

would solve my problem, but it does not work. Does anybody have an idea how to solve my problem? Any help would be greatly appreciated!

Upvotes: 2

Views: 841

Answers (1)

Divakar
Divakar

Reputation: 221504

isnan(data2winner(3,:)) gives you a logical array of ones and zeros where NaNs are present or not respectively in the third row. Then, you were needed to index into the third row of the input matrix and select the columns that had NaNs (ones from isnan output) and set them to 0.5. Your mistake was indexing into the complete matrix instead of only the third row.

Thus, this should do it -

row_id = 3;
data2winner(row_id,isnan(data2winner(row_id,:))) = 0.5

Upvotes: 3

Related Questions