#IN vs = (equals)
#IN and #NOTIN treat missing values differently from the way = and <> (and their variants: <, <=, >=, and >) do. Consider two records containing the database field X: the first record has the value 42 inX; the second record is missing a value for X (i.e., in ProAdmin expressions, its value is #MV). The following table summarizes the behavior of = and <>:
X | X=42 | X=X | X=#MV | X<>42 | X<>X | X<>#MV |
42 | 1 | 1 | 0 | 0 | 0 | 1 |
#MV | #MV | #MV | 1 | #MV | #MV | 0 |
The explanation for the behavior is this: = and <> generally act like + and *, producing #MV in the result if either argument contains a missing value. However, if one of the arguments is #MV itself (as opposed to a database field containing #MV), the operator simply tests whether the other argument is missing. (And note that = and <> are commutative, so X=#MV is the same as #MV=X.)
#IN and #NOTIN, on the other hand, treat #MV exactly like any other number, regardless of whether it occurs in the left or right argument. The result of #IN and #NOTIN is always 1’s and 0’s, never #MV.
X | X #IN 42 | X #IN X | X #IN #MV | X #NOTIN 42 | X #NOTIN X | X #NOTIN #MV |
42 | 1 | 1 | 0 | 0 | 0 | 1 |
#MV | 0 | 1 | 1 | 1 | 0 | 0 |