In an Oracle table, if a column in a row has no value, then the column is said to be null, or to contain null. Nulls can appear in columns of any datatype that are not restricted by NOT NULL or PRIMARY KEY integrity constraints. Use a null when the actual value is not known or when a value would not be meaningful.
Do not use null to represent a value of zero, because they are not equivalent. (Oracle currently treats a character value with a length of zero as null. However, this may not continue to be true in future releases, and Oracle recommends that you do not treat empty strings the same as nulls.) Any arithmetic expression containing a null always evaluates to null. For example, null added to 10 is null. In fact, all operators (except concatenation) return null when given a null operand.
NULL Comparison in Oracle
In Oracle, NULL means the lack of a value, sometimes no value is equivalent in certain circumstances to a specific value (zero for integers, for instance), but Oracle cannot make that determination generally.
So if x is a specific value, it cannot equal no value, this means that x=null and the other comparison operators will just always return false.
If x is null though, x=null will still return false, and the reason for this becomes clear if you think about exactly what you're asking. Suppose we have a table that represents the number of apples owned by people. Alice is in the table, but we don't know how many apples she has, so we insert ('Alice',NULL). Bob is in the table, ditto, ('Bob',NULL). Now we run the query "do Alice and Bob have the same number of apples?" Since we don't know how many apples either of them has, it is impossible to answer the question, so the comparison Alice=Bob will return NULL or FALSE depending how exactly you do it. Now, if not knowing how many apples someone has is equivalent in a particular scenario to them having zero apples (perhaps a better example is a salary/commission table - no commission will in most cases be equivalent to zero commission, although we might still want to distinguish between 0 and NULL so that we don't accuse non-commissioned people of having a really crappy month), then the comparison NVL(Alice,0)=NVL(Bob,0) will work and will return TRUE, because we are now comparing specific values, not NULLs.
The reason you shouldn't use equality for testing NULLs is that it's easy to get confused; if you have if (x=null) do_A else do_B, then do_B will be done because the expression evaluates FALSE. It would therefore seem logical therefore that if we replace x=null with x!=null then do_A should be done. It won't, because the comparison still returns FALSE and do_B will still be done. But if we have if (x is null) instead, then if x is not null do_B will be done, and if we replace the comparison with if (x is not null) then do_A will be done as expected.
Sometimes you'll want to select where col_A = col_B, and want the results to include where col_A and col_B are both NULL; they are not, because NULL=NULL is FALSE as explained above. You could use NVL here, but functions kill indices and there may not be a spare value that can be used, so the correct where clause would be "col_A=col_B or col_A is null and col_B is null" - somewhat clunky but crystal clear in its meaning.