invalid input syntax for type numeric

@phuctran
Agreed in general, but it's not always possible or  most effective. I would suggest that the OP consider putting 0 instead of NULL as that is the most logical value to be placed in that field. Putting NULL in that field will make your coding and SQL later more complicated as it will have to check for NULL values before doing calculations on that field.

I cannot say what number should you put in that field but you should not put null there even though the field allows you to do so.  For example, if that field is always positive (high, length, hours ...) then you can put a negative number instead of null.  Depends on your situation.  As geowrian said, 0 is a good one.

geowrian did give an example of how to perform an INSERT with a NULL; or as has been pointed out in your other question, you can just drop that column from the list of column names and don't submit the "" value to it.
  INSERT INTO ediprice_a.testing(studentid) VALUES ('beb123');

As to the issue of NULL vs. 0, as you show in your example data file, there isn't a value for the mark, so don't force it.  Yes, there are many good arguments for why NULLs should be avoided in a database.  But there are also many uses for them.  And arguments about NULLs making the SQL coding harder are just an excuse for not wanting to think about or cope with what they may mean to the underlying model.

As an example, let's say we want a table for recording patient visits to a doctor's office.  It could have a schema like this (just trying to capture enough for my example --- not intended to be complete).
Visit:  patient_id, visit_id, height, weight, arrival, departure, notes

Every time a patient visits, they are expected to have their height and weight measured.  But there may be cases where weight isn't measured: the scale is currently broken, a new assistant forgets to do it, or the patient has a stroke on the way to the scale.   (It's not likely, but it could happen.)  Do we enter "0" for the weight, or leave it empty (NULL)?  I would say we leave it empty, because that easily conveys that "we do not know what the weight is."  If we put in 0 just so we don't have a NULL, we are opening ourselves to a different coding issue which may be hidden and thus easily get missed.

Say we have a report generated to get the minimum, maximum and average weights of patients.  With the 0's in there, it gives us a skewed average and a wrong minimum.  "So just check that the weight is not zero" could be the answer.  My counter-answer is "and how is this any different than checking that the value is not NULL?"

The point is you have to think critically about your model.  If a value should never be empty, not even temporarily, then you make the column NOT NULL and you go on.  But if there are valid reasons with your data where a value could be unknown, then you let the NULLs be added to the database and deal with them, because that is just how your data lays out.  Yes, "amount_paid' does tend to suggest that it should always have a value.  But if you just put in "0" when you don't know the value, how do you then tell when something was given away for free (also "0")?  If you instead enter "-1" for NULL, how do you then tell when you paid to have some person take the item, rather than that person paying you (opposite transaction means negative)?