Building datastore on dataset with missing values

When I use readtable on my dataset that has NAN values, the readtable command converts the missing values to numeric NANs. So that works fine, but when I build a datastore on my dataset to use tall arrays, then I get the error from MATLAB that the missing values are not numerics. How can I fix that?

Thanks

Answer 1

[Jayanti]

Hi Mohsen,

When creating a datastore, you can handle missing values by using the "TreatAsMissing" and "MissingValue" name-value arguments. To replace every instance of "NA" with a 0 in the imported data specify the value of "TreatAsMissing" as "NA" and the value of "MissingValue" as 0.

The following line of code demonstrates this:

ds = datastore("airlinesmall.csv","TreatAsMissing","NA",...
               "MissingValue",0);

You can also refer the following documentation link:

https://www.mathworks.com/help/matlab/ref/datastore.html#:~:text=example-,Examples,-collapse%20all