C SUBM 10
C ..................................................................SUBM 20
C SUBM 30
C SUBROUTINE SUBMX SUBM 40
C SUBM 50
C PURPOSE SUBM 60
C BASED ON VECTOR S DERIVED FROM SUBROUTINE SUBST OR ABSNT, SUBM 70
C THIS SUBROUTINE COPIES FROM A LARGER MATRIX OF OBSERVATION SUBM 80
C DATA A SUBSET MATRIX OF THOSE OBSERVATIONS WHICH HAVE SUBM 90
C SATISFIED CERTAIN CONDITION. THIS SUBROUTINE IS NORMALLY SUBM 100
C USED PRIOR TO STATISTICAL ANALYSES (E.G., MULTIPLE REGRES- SUBM 110
C SION, FACTOR ANALYSIS). SUBM 120
C SUBM 130
C USAGE SUBM 140
C CALL SUBMX (A,D,S,NO,NV,N) SUBM 150
C SUBM 160
C DESCRIPTION OF PARAMETERS SUBM 170
C A - INPUT MATRIX OF OBSERVATIONS, NO BY NV. SUBM 180
C D - OUTPUT MATRIX OF OBSERVATIONS, N BY NV. SUBM 190
C S - INPUT VECTOR OF LENGTH NO CONTAINING THE CODES DERIVED SUBM 200
C FROM SUBROUTINE SUBST OR ABSNT. SUBM 210
C NO - NUMBER OF OBSERVATIONS. NO MUST BE > OR = TO 1. SUBM 220
C NV - NUMBER OF VARIABLES. NV MUST BE > OR = TO 1. SUBM 230
C N - OUTPUT VARIABLE CONTAINING THE NUMBER OF NON-ZERO CODESSUBM 240
C IN VECTOR S. SUBM 250
C SUBM 260
C REMARKS SUBM 270
C MATRIX D CAN BE IN THE SAME LOCATION AS MATRIX A. SUBM 280
C SUBM 290
C SUBROUTINES AND FUNCTION SUBPROGRAMS REQUIRED SUBM 300
C NONE SUBM 310
C SUBM 320
C METHOD SUBM 330
C IF S(I) CONTAINS A NON-ZERO CODE, I-TH OBSERVATION IS SUBM 340
C COPIED FROM THE INPUT MATRIX TO THE OUTPUT MATRIX. SUBM 350
C SUBM 360
C ..................................................................SUBM 370
C SUBM 380
SUBROUTINE SUBMX (A,D,S,NO,NV,N) SUBM 390
DIMENSION A(1),D(1),S(1) SUBM 400
C SUBM 410
L=0 SUBM 420
LL=0 SUBM 430
DO 20 J=1,NV SUBM 440
DO 15 I=1,NO SUBM 450
L=L+1 SUBM 460
IF(S(I)) 15, 15, 10 SUBM 470
10 LL=LL+1 SUBM 480
D(LL)=A(L) SUBM 490
15 CONTINUE SUBM 500
20 CONTINUE SUBM 510
C SUBM 520
C COUNT NON-ZERO CODES IN VECTOR S SUBM 530
C SUBM 540
N=0 SUBM 550
DO 30 I=1,NO SUBM 560
IF(S(I)) 30, 30, 25 SUBM 570
25 N=N+1 SUBM 580
30 CONTINUE SUBM 590
C SUBM 600
RETURN SUBM 610
END SUBM 620