Replacing certain values in table with NaN based on different conditions
123 views (last 30 days)
Show older comments
Hello everyone,
I'm basically new to Matlab and seem to be at a loss with my current problem - so any help would be greatly appreciated.
I have some data stored in a table called T that is 724 rows x 365 variables in size.
Now for some variables, I need to exclude extreme values based on different conditions before continuing on with my statistical analysis. But since these don't warrant excluding the whole dataset (row), I would like to replace those values with NaN in my table T.
As of now, this concerns only the variable in column 129, were values should be replaced with NaN if they are equal to or less than 3.1006 as well as column 131 were values should be replaced with NaN if they are equal to or less than 4.8673.
I have tried to get this to work for hours with several variations of logical indexing and am now stuck at this:
extremevalues = (T{:,129} <= 3.1006) | (T{:,131} <= 4.8673);
T(extremevalues) = NaN;
This does not give me any errors - which is the farthest I've come so far - however, it does not replace those extreme value in the respective columns with NaNs (there are no NaNs to be seen in either one of the specified columns in T, extremevalues however gives me several "1"s when quickly scrolling through it).
I've also tried substituting T{:, variablenr} with T.variablename - that didn't change anything either though.
I feel like the solution should be pretty easy, but I can't seem to find it (or am simply not experienced enough to find it).
Thanks very much in advance!
0 Comments
Answers (2)
darova
on 16 Aug 2021
Maybe you need to convert your data into matrix first
A = randi(10,5,4)
ind = A(:,1) < 5; % have this error when using 'table'
A(ind,:) = nan
Ive J
on 17 Aug 2021
For table, you need to directly access the variable names/column indices. so, T(extremevalues) = NaN; won't work.
This may help:
t = splitvars(table(reshape(1:20, 5, 4)))
% set values below 8 in column 2 to nan
colN = 2;
cutoff = 8;
t.(colN)(t.(colN) < cutoff) = nan
See Also
Categories
Find more on NaNs in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!