Appendix I. Control Record Formats
Records marked with * may be continued.
The record name, e.g., "INDX", is not repeated on
continuation(s).
FILE record
(FILE) (A4,4X,A72)
Field No. Value Function
1 NULL no file stream
72 chars name of file stream
SUPER record
(SUPR) (A4,4X,I4,I8,I4)
Field No. Value Function
1 1-9999 Number of problems in the superproblem
2 2-9999 Number of iterations of the superproblem.
3 0 Input information will be printed for first problem only
1 Input information will be printed for all problems
PROBLEM record
(PROB) (A4,4X,A72)
Field No. Value Function
1 72 chars problem heading
DATA record
(DATA) (A4,4X,18I4)
Field No. Value Function
1 0 or blank data set is embedded in the control stream
1 data set is in a separate file
-1 re-use the data set from the previous problem.
2 0 or blank FORTRAN unit not to be rewound
1 FORTRAN unit to be rewound
3 0 data set to be read to FINISH record or end of file
1-9999 no. of data records (low-order digits)
4 1-20 no. of data items per data record
5 0 not data checkout
1 data checkout only
6 0-9999 no. of data records (high-order digits)
The no. of data records is Field 6 * 10000 + Field 3.
When Field 6 is 0 or blank, this is simply Field 3
ITEM record
(ITEM) (A4,4X,18I4)
Field No. Value Function
1 0-20 index of ID data item
2 1-20 index of DV data item
3 0-20 index of MDV data item
4 0-20 no. of data item indices in INDXS
5 0 no user-supplied labels.
1 user-supplied labels.
6 0 standard labels PRED,RES and WRES used.
1 nonstandard labels used.
7 0-20 index of L2 data item
8 0-20 index of first data item specified in CONTR record
9 0-20 index of second data item specified in CONTR record
10 0-20 index of third data item specified in CONTR record
11 0-50 no. of user-supplied labels for tables, scatters
12 0-20 index of MRG_ data item
13 0-20 index of RAW_ data item
14 0-20 no. of items on OMIT record
15 0-20 index of RPT_ data item
INDEX record
(INDX)* (A4,4X,18I4)
Field No. Value Function
1 1-20 1st element of INDXS
2 1-20 2nd element of INDXS
etc.
LABEL record
(LABL)* (A4,4X,9(A4,4X))
Field No. Value Function
1 4 chars label of 1st data item
2 4 chars label of 2nd data item
etc.
m 4 chars label of last data item
m+1 4 chars label for PRED (if ITEM(6)=1)
m+2 4 chars label for RES (if ITEM(6)=1)
m+3 4 chars label for WRES (if ITEM(6)=1)
m+p+1 4 chars label for 1st variable in NMPRD4†
m+p+2 4 chars label for 2nd variable in NMPRD4†
etc.
m+p+q 4 chars label for last displayed variable in
NMPRD4
Note
m=no. of data items per data rec.=DATA(4)
p=3 if non-standard labels for PRED, RES, WRES (ITEM(6)=1)
p=0 otherwise
q=no. of user supplied labels for tables, scatters=ITEM(11)
† Blank if this variable is not displayed
OMIT record
(OMIT)* (A4,4X,18I4)
Field No. Value Function
1 4 chars no. of 1st data item omitted from template
matching
2 4 chars no. of 2nd data item omitted from template
matching
etc.
FORMAT record
(FORM) (A4,4X,A72/A80)
Field No. Value Function
1 80 chars format specification
(field begins on first continuation record)
FIND record
(FIND) (A4,4X,18I4)
Field No. Value Function
1 0
2 0
3 0 No Model specification file (MSFI)
1 A Model specification file (MSFI) is to be read.
4 0 estimate on file not to be rescaled.
1 estimate on file to be rescaled.
5 0 No ONLYREAD option
1 ONLYREAD option
initial STRUCTURE record
(STRC) (A4,4X,18I4)
Field No. Value Function
1 0-70 length of THETA
2 0-70 dimension of OMEGA
3 0-70 dimension of SIGMA
4 blank
5 blank
6 0 or blank OMEGA constrained with a block set partition
1 OMEGA constrained to be diagonal
7 0 or blank only if field 6 has value 1
1-70 number of block sets for OMEGA
If the dimension of SIGMA is 0,
the following fields may be ignored.
8 0 or blank SIGMA constrained with a block set partition
1 SIGMA constrained to be diagonal
9 0 or blank SIGMA only if field 8 has value 1
1-70 number of block sets for SIGMA
10 blank
11 blank
12 0 or blank default THETA boundary test
1 No default THETA boundary test
13 0 or blank default OMEGA boundary test
1 No default OMEGA boundary test
14 0 or blank default SIGMA boundary test
1 No default SIGMA boundary test
STRUCTURE record for OMEGA or
SIGMA (STRC)* (A4,4X,18I4)
Field No. Value Function
1 1-70 size of 1st. block set
2 1-70 dimension of blocks in 1st. block set
3 1-70 size of 2nd. block set
4 1-70 dimension of blocks in 2nd. block set
etc.
THETA CONSTRAINT record
(THCN) (A4,4X,18I4)
Field No. Value Function
1 0 or blank THETA unconstrained
1 THETA constrained
2 0 or blank use default size of initial. est. search
1-9999 no. of points to be examined during initial est.
search.
3 0 or blank ABORT if PRED sets error return code to 1
during search
1 NOABORT - Ignore PRED error return code during search
2 NOABORTFIRST - Same, even with first values.
THETA record
(THTA)* (A4,4X,9A8)
Field No. Value Function
1 initial est. of θ 1
(blank if NONMEM is to obtain the inital est.)
2 initial est. of θ 2
(blank if NONMEM is to obtain the inital est.)
etc.
LOWER BOUND record
(LOWR)* (A4,4X,9A8)
Field No. Value Function
1 lower bound for θ 1
2 lower bound for θ 2
etc.
UPPER BOUND record
(UPPR)* (A4,4X,9A8)
Field No. Value Function
1 upper bound for θ 1
2 upper bound for θ 2
etc.
DIAGONAL record
(DIAG)* (A4,3X,A1,9A8)
for OMEGA or SIGMA
Field No. Value Function
Pos. 1 blank Not fixed.
1 Fixed.
2 NONMEM is to obtain the inital estimate(s).
1 initial est. of (1,1) element of matrix
2 initial est. of (2,2) element of matrix
etc.
BLOCK SET record
(BLST)* (A4,3X,A1,9A8)
for OMEGA or SIGMA
Field No. Value Function
Pos. 1 blank Not fixed.
1 Fixed.
2 NONMEM is to obtain the inital estimate(s).
1 initial est. of (1,1) element of matrix
2 initial est. of (1,2) element of matrix
etc.
use symmetric enumeration
SIMLUATION record
(SIML) (A4,4X,18I4)
Field No. Value Function
1 0 or blank Simulation Step implemented
1 Simulation Step not implemented
If the value is 1, the
subsequent fields may be ignored.
2 1-10 no. of random sources (SORC records)
3 0 eta (eps) changes with each record
1 eta (eps) changes with new ind.rec. (L2 rec) (NEW)
4 0-9999 no. of subproblems
5 0 compute objective function and other steps
1 only the simulation step
6 0 or blank no partial derivatives from PRED needed
1 PRED should compute 1st. derivatives (REQUESTFIRST)
2 PRED should compute 2nd. derivatives (REQUESTSECOND)
7 0 or blank simulated observation is Y or F (PREDICTION)
1 simulated observation is DV (NOPREDICTION)
8 0 or blank Use inital ests. (TRUE=INITIAL)
1 with MSFI, use final ests. (TRUE=FINAL)
2 use values in NMPR16 (TRUE=PRIOR)
SOURCE record (SORC) (A4,4X,2A12,I4)
ESTIMATION record
(ESTM) (A4,4X,18I4)
Field No. Value Function
1 0 or blank Estimation Step implemented
1 Estimation Step not implemented
If the value is 1, the
subsequent fields may be ignored.
2 0-9999 maximum no. of function. evaluations (low-order
digits)
-1 Reuse the value from the previous run (with MSFI)
3 1-8 number of significant figs. required in final est.
4 0 or blank no summarization of iterations
n>0 every nth iteration summarized
5 0 or blank no second search (REPEAT)
1 second search (REPEAT) implemented
6 0 or blank MSF not output
1 MSF output
7 0 or blank First order (FO) method
1 Conditional method (METHOD=COND)
8 0 or blank No POSTHOC etas are to be estimated.
1 POSTHOC etas are to be estimated.
9 0 or blank Etas are 0 for comp. of intraind. error
(NOINTERACTION)
1 Nonzero etas for comp. of intraind. error INTERACTION
10 0 or blank Do not use Laplacian method.
1 Laplacian method is to be used.
11 0 or blank ABORT if PRED sets error return code to 1
1 NOABORT - Attempt theta-recovery when PRED error code 1.
12 0 or blank Faster method of compuation (NOSLOW)
1 Slower method of computation (SLOW)
2 Slower method of computation (SLOW=2); for Stieltjes
13 0 or blank avg. cond. est. of etas unconstrained
(NOCENTER)
1 avg. cond. est. of etas constrained close to 0. (CENTER)
14 0 or blank First-order model not used (NOFO)
1 First-order model used with METHOD=1 CENTERING (FO)
15 0 or blank Second eta-derivs. computed by PRED
(NONUMERICAL)
1 Second eta-derivs. for Laplacian to be obtained
numerically.
16 0 or blank Y or F (with user-supplied code) is a
prediction.
1 Y or F is a LIKELIHOOD.
2 Y or F is a -2LOGLIKELIHOOD
17 0 or blank Not the Hybrid method
1-99 no. of etas fixed to zero by ZERO recs. (Hybrid method)
18 0 or blank Not the Stieltjes method.
1 Stieltjes method; no GRID option.
2 Stieltjes method; GRID was specified.
ESTIMATION rec. continuation
rec. ( ) (A4,4X,18I4)
Field No. Value Function
1 0 or blank Required if estimation step is omitted,
otherwise:
0 or blank The REPEAT2 option is not coded; same as
NOREPEAT2
1 REPEAT2 (with Stieltjes)
2 0 or blank No ETABARCHECK.
1 ETABARCHECK option is coded.
3 0 or blank. Sum contrib. to obj. func. in data set order.
1 Sort contrib. to obj. func. prior to sum (SORT)
4 0-9999 maximum no. of function evaluations (high-order
digits)
The no. of func. evals. is Field 4 * 10000 + low-order
When Field 4 is 0 or blank, this is simply low-order
ZERO record
(ZERO)* (A4,4X,18I4)
Field No. Value Function
1 0 conditional estimate for eta(1)
1 eta(1) is fixed to 0 (HYBRID method)
2 0 conditional estimate for eta(2)
1 eta(2) is fixed to 0 (HYBRID method)
etc.
GRID record
(GRID) (A4,4X,9A8)
Field No. Value Function
1 nr as specified in GRID=(nr,ns,r0,r1)
2 ns as specified in GRID=(nr,ns,r0,r1)
3 r0 as specified in GRID=(nr,ns,r0,r1)
4 r1 as specified in GRID=(nr,ns,r0,r1)
NONPARAMETRIC record
(NONP) (A4,4X,18I4)
Field No. Value Function
1 0 or blank Nonparametric step implemented conditionally
1 Nonparametric step implemented unconditionally
2 0 or blank use nonparametric estimate from input MSF
1 recompute nonparametric estimate
3 0 or blank obtain marginal cumulatives
1 compute conditional nonpar. etas (CNPE ETAS)
4 0 or blank no model specification file is output
1 a model specification file is output
COVARIANCE record
(COVR) (A4,4X,18I4)
Field No. Value Function
1 0 or blank Covariance Step conditionally implemented
(1) (Covariance Step unconditionally implemented - obsolete)
2 Covariance Step not implemented
2 0 or blank covariance matrix set to (R inverse) S (R
inverse)
1 covariance matrix set to R inverse
2 covariance matrix set to S inverse
3 0 or blank neither R nor S printed.
1 R matrix printed
2 S matrix printed
3 both R and S printed
4 0 or blank eigenvalues not printed
1 eigenvalues printed.
5 0 or blank default computation.
1 Special computation with a recursive PRED subroutine.
6 0 or blank Print Covariance Step arrays in normal format.
1 Print Covariance Step arrays in compressed format.
7 1
8 0 or blank
9 0 or blank Normal method of computation
1 Slower method of computation (SLOW)
initial TABLE record
(TABL) (A4,4X,18I4)
Field No. Value Function
1 0 or blank Table Step conditionally implemented
1 Table Step unconditionally implemented
2 Table Step not implemented
If the value is 2, the next
field may be ignored, and there should not appear
any individual TABLE records.
2 1-10 number of tables
individual TABLE record
(TABL) (A4,3X,I1,18I4)
Field No. Value Function
Pos. 1 blank no option record.
1 an option record follows.
(only if at least one item on the option rec. is non-blank)
1 0-50 number of selected data item types
2 1-999 index of 1st selected data item type
3 0-8 sort code for data items of 1st selected type
4 1-999 index of 2nd selected data item type
5 0-8 sort code for data items of 2nd selected type
etc.
individual TABLE rec. contin.
rec. ( )* (A4,4X,18I4)
(as needed)
individual TABLE record option
rec. ( ) (A4,4X,18I4)
Field No. Value Function
1 blank Every data record appears in the table.
1 Only the first data rec. from each ind. rec. (FIRSTONLY)
2 1 With TABLE file, no printed table (NOPRINT)
2 With TABLE file, printed table appears in the NONMEM
output.
3 blank Normal header lines appear in the TABLE file.
1 Only one header in the TABLE file (ONEHEADER)
2 No headers are included in the TABLE file (NOHEADER)
4 blank The TABLE file is opened and is positioned at the
start.
1 The TABLE file is positioned at the end (FORWARD)
5 blank DV, PRED, RES, WRES appear automatically
1 DV, PRED, RES, WRES do not appear unless listed
(NOAPPEND)
initial SCATTERPLOT record
(SCAT) (A4,4X,18I4)
Field No. Value Function
1 0 or blank Scatterplot Step conditionally implemented
1 Scatterplot Step unconditionally implemented
2 Scatterplot Step not implemented
If the value is 2, the next
field may be ignored, and there should not appear
any individual SCATTERPLOT records.
2 1-20 number of families
individual SCATTERPLOT record
(SCAT) (A4,4X,6I4,2I8,4I4,16X)
Field No. Value Function
1 1-23 index of data items plotted on abcsissa axis
2 1-23 index of data items plotted on ordinate axis
3 0 or blank a single scatterplot
1 a one-way partitioned scatterplot
2 a two-way partitioned scatterplot
If the value of field 3 is 0 or
blank, the next two fields should be ignored.
4 1-23 index of 1st separator
If the value of field 3 is 1,
the next field should be ignored.
5 1-23 index of 2nd separator
6 0 or blank no unit slope line appears
1 unit slope line appears
7 0-99999999 no. of the first data rec. for the scatter
(FROM)
8 0-99999999 no. of the last data rec. for the scatter (TO)
9 0 or blank a line through zero on the ordinate axis if
appropriate.
1 a line through zero on the ordinate axis. (ORD0)
-1 no line through zero on the ordinate axis.
10 0 or blank a line through zero on the abscissa axis if
appropriate.
1 a line through zero on the abscissa axis. (ABS0)
-1 no line through zero on the abscissa axis.
11 0 or blank Every data record appears in the scatter.
1 Only the first data rec. from each ind. rec. (FIRSTONLY)
12 0 or blank Every data record appears in the scatter
1 Only data records with MDV=0 (OBSONLY).