How to get NaN if the data is not in proper input format or missing?

I am running a function over each table rows. It is a big table and some of the entries are 'NaN' (class: 'double') or 'none' (class: 'char'). There may some missing entries too. I want to get 'NaN' as output if any of the table entries is not a valid entry. How should I approach it? I have uploaded the table in attachment. My function is as follows.

function [logical_out, reaction_time, average_position] = new_table2(name,date,Trial)

[dpb]

What does this function actually do/use to get the outputs? rowfun wiith the table is probably the tool to use; in general the result of an operation with MATLAB on numeric NaN inputs is NaN output; you'll get that automagically without even trying. Similarly for char() data although I'd strongly recommend the cellstr or convert to categorical and then the <missing> indicator will also come along.

If there are cases where one variable is missing but not all in the given row and it's needed to have all inputs, then the all function will let you test for an array or logical tests for any of the needed variables will let you know about which are/are not valid records.

Just give 'er a go...come back with specific Q? when you run into a roadblock -- but then show us the code itself; we can't do anything with just the function definition line.

[Atanu]

Thank you. Let me try.

[Atanu]

You are right. All the NaN entries for 'playstarttrialtone' and 'presentcost' gave me NaN as output without doing anything.

By the way, I have converted all the columns from cellstring to double (e.g., playstarttrialtone and presentcost) or string ('subjectid') as I thought would be appropriate. But the bad entries in the subjectid column is specified as 'none' in the table. So they are still giving me outputs.

Answer 1

[Walter Roberson]

https://www.mathworks.com/help/matlab/ref/standardizemissing.html

standardize missing. Then rmmissing https://www.mathworks.com/help/matlab/ref/rmmissing.html with the second output to show which rows were removed. Run the function on the reduced data, and then inject the results into a larger vector of nan values to match up to the original rows.

[Atanu]

Thank you! Let me try this.