Any array must be rectangular; char() arrays are no different. The array must be as wide (have as many columns) as the longest char() string you wish it to hold and all rows must be padded to that length if they are shorter than the maximum length. You can either pre- or postpend blanks, but you can't not fill all the space with something. If you try to fool mother Matlab, you'll either get a linear string if it can interpret the comand to do so or perhaps what appears as "gibberish" if you try to place values outside the range of printing ASCII characters into a char() array as you've done above.
That's why there are cellstr() and/or strings to handle variable length strings as entities, not as just arrays of bytes of type char.
What is the end result to be made of this variable in the resulting table and why do you think a char() representation is "more better" than string() or cellstr()? Frankly, it looks like perhaps a categorical variable might be more suitable than either, but don't know the objective so can't say for certain lacking additional input.
Best Answer