*This Q on Stackoverflow - http://stackoverflow.com/q/22643216/604456.

*All this takes is a list of variables & prints the intersection of the lists and cardinality.
*Does not count repeat symbols in the lists.
*Uses strings in matrix - so variable names will be truncated to first 8 values.
DEFINE !InterSet (!POSITIONAL = !CMDEND)
OMS /SELECT ALL EXCEPT = [WARNINGS] /TAG = "NoCrap" /DESTINATION VIEWER = NO.
DATASET COPY XTempX.
DATASET ACTIVATE XTempX.
!LET !Total = !NULL
!DO !I !IN (!1)
  !LET !Total = !CONCAT(!Total," ")
  !LET !E = !I
!DOEND
!LET !Ln = !LENGTH(!Total)
!LET !B = !HEAD(!1)
MATCH FILES FILE = * /KEEP !1.
VARSTOCASES /MAKE V FROM !B TO !E /INDEX OrigVar (V).
SORT CASES BY V OrigVar.
CASESTOVARS /ID = V /VIND ROOT = "D" /INDEX = OrigVar. 
!DO !I !IN (!1)
RENAME VARIABLES (!CONCAT("D",!I) = !I).
!DOEND
*This has to be a set of symbols not contained in the.
*Original variable name anywhere.
!LET !pars = "0x_x0"
!DO !I !IN (!1)
!DO !J !IN (!1)
  COMPUTE !CONCAT(!I,!pars,!J) = !I*!J.
!DOEND
!DOEND
COMPUTE Order = $casenum.
DO REPEAT D = !CONCAT(!B,!pars,!B) TO !CONCAT(!E,!pars,!E).
  IF (Order<>1) D = LAG(D) + D.
END REPEAT.
SORT CASES BY Order (D).
SELECT IF ($casenum = 1).
FLIP VARIABLES = !CONCAT(!B,!pars,!B) TO !CONCAT(!E,!pars,!E).
COMPUTE CASE_LBL = CHAR.SUBSTR(CASE_LBL,CHAR.INDEX(CASE_LBL,!QUOTE(!pars))+!LENGTH(!pars)).
OMSEND TAG = "nocrap".
MATRIX.
GET I /FILE = * /VARIABLE = var001.
GET V /FILE = * /VARIABLE = CASE_LBL.
COMPUTE I2 = RESHAPE(I,!Ln,!Ln).
COMPUTE V2 = V(1:!Ln).
PRINT I2 /TITLE = "Intersection of Lists" /RNAMES =V2 /CNAMES = V2.
COMPUTE V3 = {"Var";V2}.
SAVE {V2,I2} /OUTFILE = * /NAMES V3 /STRINGS Var.
END MATRIX.
DATASET NAME InterSet.
FORMATS !B TO !E (F3.0).
DATASET CLOSE XTempX.
!ENDDEFINE.

*Example use.
 * DATASET CLOSE ALL.
 * OUTPUT CLOSE ALL.
 * DATA LIST FREE / x_1 (A1) y_1 (A5) z_1 (A8).
 * BEGIN DATA
 *    b    #
 g    #    i
 #    *    l
 +    k    ""
 ""   m    ""
END DATA.
 * DATASET NAME Test.
*InterSet x_1 y_1 z_1 .