How to remove NaN from data read from excel. Nan values only not rows or columns

I have read 4 columns of data from an excel file. some columns have more data than others. Nan values have populated the blank cells. However I need to plot each column of data and the Nan values are preventing this. How do i either import the data without the nan values or once i have imported it how do I get rid of the nan values? I cant use code that removes rows or columns as otherwise i will lose valuable data. this some code i have been trying:

[~,sheet_name]=xlsfinfo('barchart_F1.xlsx');
barchart=xlsread('barchart_F1.xlsx');
barchart = importdata('barchart_F1.xlsx');
barchart=rmmissing(barchart); 

[Star Strider]

First, use readmatrix or readtable to import the file.

What is to be done with the data once read and imported?

[suzanne o'callaghan]

the data in each column is to be printed on the one plot.

[suzanne o'callaghan]

hi star, i have tried readmatrix and readtable and NANs are still populating the table

[Johannes Hougaard]

In MATLAB your array is not made up of individual values/cells/columns/rows. MATLAB is focused around matrices not around values, whereas in excel each cell is separate and has no relation to neighboring cells.

If you wish to mimic that in MATLAB each cell from excel should be it's own variable (and you don't want that for the data you've got)

The matrix you read is of size 18x4 you can't have a different number of rows in different columns.

Answer 1

[chiyaan chandan]

clc
clear all
clf
%%
disp ('Select User input excel file ');
[filename, pathname] = uigetfile('*.*');
xfile1 = fullfile(pathname, filename);
Data=  xlsread(xfile1);
Data(isnan(Data))=0;                   % replace the NaN with 0 
%%
[m,n]=size(Data);
for i=1:n
    figure(i)
    bar(Data(:,i))
end

once we replcae the NaN with 0 at the end of the coloumns then we plaot the barchart for for each coloumns in a different figures or we can plot in single barchart.

[Johannes Hougaard]

NaN values are perfectly fine for making the bar chart, there is no need to replace them with 0.

The graphs will be identical whether you keep the NaN (missing values) or replace them with 0 (actual values) but all statistics (such as the mean, standard deviation etc.) will be wrong when replacing zeros.

tmp = rand(18,4)*2 + 7;
tmp(14:end,4) = missing;
tmp(7:end,1) = missing;
figure;
bar(tmp');
ah = gca;
ah.ColorOrder = jet(18);

mean_values = mean(tmp,1,"omitnan")

mean_values = 1×4

8.0884 8.1448 8.1048 8.1167

tmp(14:end,4) = 0;
tmp(7:end,1) = 0;
figure;
bar(tmp');
ah = gca;
ah.ColorOrder = jet(18);

mean_values = mean(tmp,1,"omitnan")

mean_values = 1×4

2.6961 8.1448 8.1048 5.8620