Movatterモバイル変換

Nonmem-style data set. If using 'TAPD' an 'EVID'column must contain 1 for dosing records.

A numerical vector with simulation times. Can also bea data.frame in which case it must contain a 'TIME' column andis merged with 'data'.

A numerical vector with simulation times, relative toprevious dose. When this is used, 'data' must contain rowswith 'EVID=1' events and a 'TIME' column. 'TAPD' can also be adata.frame in which case it must contain a 'TAPD' column andis merged with 'data'.

The compartment in which to insert the EVID=2records. Required if 'CMT' is a column in 'data'. If longerthan one, the records will be repeated in all the specifiedcompartments. If a data.frame, covariates can be specified.

The value to put in the 'EVID' column for the createdrows. Default is 2 but 0 may be prefered even for simulation.

Optionally provide a single value to be assigned to the'DV' column. The default is to assign nothing which willresult in 'NA' as samples are stacked ('rbind') with'data'. If you assign a different value in 'DV', the defaultvalue of 'EVID' changes to '0', and 'MDV' will be '0' insteadof '1'. An example where this is useful is when generatingdatasets for '$DESIGN' where 'DV=0' is often used.

The name of the column in 'data' that holds theunique subject identifier. Currently, this is needed to benon-'NULL'.

Only relevant - and likely not needed -if data contains ADDL and II columns. If those columns areincluded, 'NMaddSamples()' will use 'NMdata::NMexpanDoses()'to evaluate the time of each dose. Other than the 'data'argument, 'NMaddSamples()' relies on the default'NMexpanDoses()' argument values. If this is insufficient, youcan specify other argument values in a list, or you can call'NMdata::NMexpanDoses()' manually before calling'NMaddSamples()'.

If 'TRUE' (default), events are reduced to uniquetime points before insertion. Sometimes, it's easier tocombine sequences of time points that overlap (maybe across'TIME' and 'TAPD'), and let 'NMaddSamples()' clean them. Ifyou want to keep your duplicated events, use 'unique=FALSE'.

IfTIME and/or 'TAPD' are 'data.frame's andcontain other columns than 'TIME' and/or 'TAPD', those will bydefault follow the 'TIME'/'TAPD' records. Think of them asrecord-level variables, like 'VISIT'. The exception is'col.id' - if the subject identifier is present, it will bemerged by. If additional columns should be used to merge by,you can use the 'by' argument. This is useful to generatedifferentiated sampling schemes for subsets of subjects (likeregimen="SAD" and regimen="MAD"). If no columns in 'TIME'and/or 'TAPD' should not be merged by, use 'by=FALSE'. You canalso specify selected 'by' variables like 'by="ID"' or'by=c("ID","regimen")' See examples.

Suppress messages? Default is 'FALSE'.

The default is to return data as a'data.frame'. Pass a function (say 'tibble::as_tibble') inas.fun to convert to something else. If data.tables arewanted, use 'as.fun="data.table"'. The default can beconfigured using 'NMdataConf()'.

Deprecated. Use 'data'.

Deprecated. Use 'TIME'.

Deprecated. Use 'by'.

The time of the dosing events. Required.

vector or data.frame with amounts amount. Required.

The event ID to use for doses. Default is to useEVID=1, but EVID might also be wanted.

Compartment number. Default is to dose into CMT=1. Use'CMT=NA' or 'CMT=NULL' to omit in result.

Number of additional dose events. Must be incombination with and consistent with II. Notice if of length1, only applied to last event in each regimen.

Dosing frequency of additional events specified in'ADDL'. See 'ADDL' too.

Infusion rate. Optional.

steady-state flag. Optional.

A list of ADDL and II that will be applied to lastdose. This may be prefered if II and ADDL depend on covariates- see examples. Optional.

Number of replications. Default is 1. If 'N=1' results intwo distinct subjects, 'N=100' will result i 200 distinctsubjects. The ID column will automatically be recoded tocontain distinct ID's.

If ADDL and II are of length 1, apply only tolast event of a regimen? The default is 'TRUE'.

Default is to denote the dosing regimens by an IDcolumn. The name of the column can be modified using thisargument. Use 'col.id=NA' to omit the column altogether. Thelatter may be wanted if repeating the regimen for a number ofsubjects after running 'NMcreateDoses()'.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

A data.table with at least 'i', 'j' and 'value'columns. See 'NMdata::NMreadExt' and the pars element returnedby that function. Must at least have columns 'i', 'j','value', 'iblock', 'blocksize', 'FIX'.

If 'TRUE', all values are printed as oneblock. If 'FALSE' (default), matrix will be separeted intoblocks based on position non-zero off-diagonalvalues. Generally speaking, for 'OMEGA' matrices (var-covmatrices for ETAs), this should be 'FALSE', and forvariance-covariance matrices (like 'THETAP'), this should be'TRUE'.

Include 'FIX' for all lines? If 'FALSE', fixing willnot be modified. Notice, 'fix=TRUE' will fix everything -individual parameters cannot be controlled. For finer controland way more features, see 'NMdata::NMwriteInits()'.

The matrix type. 'OMEGA' or 'SIGMA' - casein-sensitive. Will be used to print say '$OMEGA' in front ofeach line.

File paths to the models (control streams) to runnonmem on. See file.pattern too.

Alternatively to files, you can supply aregular expression which will be passed to list.files as thepattern argument. If this is used, use dir argument aswell. Also see data.file to only process models that use aspecific data file.

If file.pattern is used, dir is the directory to searchfor control streams in.

Use the sge queing system. Default is TRUE. Disable forquick models not to wait for the queue to run the job.

A function of the model file path to generatethe path in which to archive the input data as RDS. Set toFALSE not to archive the data.

Number of cores to use if sending to the cluster. Thiswill only be used ifmethod.execute="psn", andsge=TRUE. Default is 64.

The directory in which the data file isstored. This is normally not needed as data will be foundusing the path in the control stream. This argument may beremoved in the future since it should not be needed.

Wait for process to finish before making R consoleavailable again? This is useful if calling NMexec from afunction that needs to wait for the output of the Nonmem runto be available for further processing.

The path to the nonmem executable. Only used ifmethod.execute="direct" ormethod.execute="nmsim" (which is not default). If thisargument is not supplied, NMexec will try to run nmfe75,i.e. this has to be available in the path of the underlyingshell. The default value can be modified usingNMdata::NMdataConf, likeNMdataConf(path.nonmem="/path/to/nonmem")

Only run model(s) if control stream or dataupdated since last run?

A function of the path to the control stream('file.mod') that generates bash code to be evaluated onceNonmem is done. This can be used to automatically run agoodness-of-fit script or a simulation script after modelestimation.

How to run Nonmem. Must be one of 'psn','nmsim', or 'direct'.

• psn PSN's execute is used. This supports parallel Nonmemruns. Use thenc argument to control how many cores to usefor each job. For estimation runs, this is most likely the betterchoice, if you have PSN installed. Seedir.psn argumenttoo.

• nmsim Creates a temporary directory and runs Nonmeminside that directory before copying relevant results files backto the folder where the input control stream was. Ifsge=TRUE, the job will be submitted to a cluster, butparallel execution of the job itself is not supported. Seepath.nonmem argument too.

• direct Nonmem is called directly on the control stream. Thisis the simplest method and is the least convenient in mostcases. It does not offer parallel runs and leaves all the Nonmemoutput files next to the control streams.

See 'sge' as well.

additional options that will be passed tonmfe. It is only used when path.nonmem is available (directlyor using 'NMdataConf()'). Default is "-maxlim=2" For PSN, see'args.psn.execute'.

The directory in which to find PSNexecutables. This is only needed if these are not searchablein the system path, or if the user should want to be explicitabout where to find them (i.e. want to use a specificinstalled version of PSN).

A character string with arguments passedto execute. Default is"-model_dir_name -nm_output=coi,cor,cov,ext,phi,shk,xml".

In case method.execute="nmsim", this argumentspecifies files to be copied into the temporary directorybefore Nonmem is run. Input control stream and simulationinput data does not need to be specified.

The degree of cleaning (file removal) to do afterNonmem execution. If 'method.execute=="psn"', this is passedto PSN's 'execute'. If 'method.execute=="nmsim"' a similarbehavior is applied, even though not as granular. NMsim'sinternal method only distinguishes between 0 (no cleaning),any integer 1-4 (default, quite a bit of cleaning) and 5(remove temporary dir completely).

Before running, should existing results files bebacked up in a sub directory? If not, the files will bedeleted before running.

Suppress messages on what NMexec is doing? Default isFALSE.

Suppress terminal output from 'Nonmem'. This islikely to only work on linux/unix systems.

A charachter string, either \"windows\" or\"linux\" - case insensitive. Windows is only experimentallysupported. Default is to useSys.info()[["sysname"]].

Path to a Nonmem input control stream.

Path to Nonmem executable. You may want tocontrol this withNMdata::NMdataConf.

Files needed to run the control stream. Thiscold be a .phi file from which etas will be read. Notice,input data set will be handled automatically, you do not needto specify that.

If NULL, data will be copied into the temporarydirectory, and Nonmem will read it from there. If not NULL,dir.data must be the relative path from where Nonmem is run towhere the input data file is stored. This would be ".." if therun directory is created in a directory where the data isstored.

The degree of cleaning (file removal) to do afterNonmem execution. If 'method.execute=="psn"', this is passedto PSN's 'execute'. If 'method.execute=="nmsim"' a similarbehavior is applied, even though not as granular. NMsim'sinternal method only distinguishes between 0 (no cleaning),any integer 1-4 (default, quite a bit of cleaning) and 5(remove temporary dir completely).

A file path to read from. Normally a .mod or .lst. Seelines and text as well.

Text lines to process. This is an alternative tousing the file and text arguments.

Use this argument if the text to process is one longcharacter string, and indicate the line separator with thelinesep argument. Use only one of file, lines, and text.

The name of section to extract. Examples: "INPUT","PK", "TABLE", etc. It can also be result sections like"MINIMIZATION".

The section denoted as a string compatiblewith regular expressions. "$" (remember to escape properly)for sections in .mod files, "0" for results in .lst files.

A regular expression to capture the end of thesection. The default is to look for the next occurrence ofchar.section.

If "text", plain text lines are returned. If "idx",matching line numbers are returned. "text" is default.

Keep empty lines in output? Default isFALSE. Notice, comments are removed before empty lines arehandled if 'keep.comments=TRUE'.

Keep the section name in output (say, "$PROBLEM")Default is TRUE. It can only be FALSE, if return="text".

Default is to keep comments. If FALSE, thewill be removed.

If multiple hits, concatenate into one. This willmost often be relevant with name="TABLE". If FALSE, a listwill be returned, each element representing a table. Defaultis TRUE. So if you want to process the tables separately, youprobably want FALSE here.

If TRUE, leading and trailing are removed, andmultiplied succeeding white spaces are reduced to single whitespaces.

If asOne=FALSE, do you want the result to besimplified if only one table is found? Default is TRUE whichis desirable for interactive analysis. For programming, youprobably want FALSE.

Default is to search for exact matches of'section'. If FALSE, only the first three characters arematched. E.G., this allows "ESTIMATION" to match "ESTIMATION"or "EST".

Either mod, res or NULL. mod is for information thatis given in .mod (.lst file can be used but results section isdisregarded). If NULL, NA or empty string, everything isconsidered.

If using the text argument, use linesep to indicatehow lines should be separated.

Deprecated. See keep.empty.

Deprecated. See keep.name.

Deprecated. See keep.comments.

Deprecated. See as.one.

number of cores wanted

The file path to write the result to

Control stream path

Control stream lines if already read from file

Return the filters only or also return the remaining text in a separate object? If 'FALSE', a list with the two objects is returned.

Function to run on the tables with filters.

Path to a control stream. See 'lines' too.

A control stream as text lines. Use this or 'file'.

The section to read. Typically, "theta", "omega",or "sigma". Default is those three.

By default (whenreturn="pars", a parametertable with initial values, FIX, lower and upper bounds etc. Inmost cases, that is what is needed to derive information aboutparameter definitions. Ifreturn="all", two additionaltables are returned which can be used if the aim is to modifyand write the resulting parameters to a control stream.

See ?NMscanData

A file path to read from. Normally a .mod or .lst. Seelines also.

Text lines to process. This is an alternative tousing the file argument.

Deprecated, use 'lines'. Use this argument if the textto process is one long character string, and indicate the lineseparator with the linesep argument (handled byNMextractText). Use only one of file, lines, and text.

The name of section to extract without"$". Examples: "INPUT", "PK", "TABLE", etc. Not casesensitive.

If "text", plain text lines are returned. If "idx",matching line numbers are returned. "text" is default.

Keep empty lines in output? Default isFALSE. Notice, comments are removed before empty lines arehandled if 'keep.comments=TRUE'.

Keep the section name in output (say, "$PROBLEM")Default is FALSE. It can only be FALSE, if return="text".

Default is to keep comments. If FALSE, thewill be removed. See keep.empty too. Notice, there is no wayfor NMreadSection to keep comments and also drop lines thatonly contain comments.

If multiple hits, concatenate into one. This willmost often be relevant with name="TABLE". If FALSE, a listwill be returned, each element representing a table. Defaultis TRUE. So if you want to process the tables separately, youprobably want FALSE here.

If TRUE, leading and trailing are removed, andmultiplied succeeding white spaces are reduced to single whitespaces.

If asOne=FALSE, do you want the result to besimplified if only one section is found? Default is TRUE whichis desirable for interactive analysis. For programming, youprobably want FALSE.

Deprecated. See keep.empty.

Deprecated. See keep.name.

Deprecated. See keep.comments.

Deprecated. See as.one.

Additional arguments passed to NMextractText

Path to the simulation-specific rds file generated byNMsim, typically called 'NMsim_MetaData.rds'. Can also be atable of simulation runs as stored in 'rds' files by'NMsim'. The latter should almost never be used.

If found, check whether 'fst' file modificationtime is newer than 'rds' file. The 'fst' is generated based oninformation in ‘rds', but notice that some systems don’tpreserve the file modification times. Becasue of that,'check.time' is 'FALSE' by default.

By default, 'NMreadSim' will use information aboutthe relative path from the results table file('_MetaData.rds') to the Nonmem simulation results. If thesepaths have changed, or for other reasons this doesn't work,you can use the 'dir.sims' argument to specify where to findthe Nonmem simulation results. If an '.fst' file was alreadygenerated and is found next to the '_MetaData.rds', the pathto the Nonmem simulation results is not used.

If simulations seem to not be done yet, wait for themto finish? If not, an error will be thrown. If you choose towait, the risk is results never come. 'NMreadSim' will bewaiting for an 'lst' file. If Nonmem fails, it will normallygenerate an 'lst' file. But if 'NMTRAN' fails (checks ofcontrol stream prior to running Nonmem), the 'lst' file is notgenerated. Default is not to wait.

Turn off some messages about what is going on?Default is to report the messages.

Track progress? Default is 'TRUE' if 'quiet' isFALSE and more than one model is being read. The progresstracking is based on the number of models completed/read, notthe status of the individual models.

Skip models where results are not available?Default is 'FALSE' meaning an error will be thrown if one ormore models do not have completed results.

If results are read successfully, remove temporarysimulation results files? This can be useful after a script isdeveloped and intermediate debugging information is notneeded. It cleans up and saves significant disk space.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

Path to the simulation-specific rds file generated byNMsim, typically called 'NMsim_MetaData.rds'. Can also be atable of simulation runs as stored in 'rds' files by'NMsim'. The latter should almost never be used.

If found, check whether 'fst' file modificationtime is newer than 'rds' file. The 'fst' is generated based oninformation in ‘rds', but notice that some systems don’tpreserve the file modification times. Becasue of that,'check.time' is 'FALSE' by default.

By default, 'NMreadSim' will use information aboutthe relative path from the results table file('_MetaData.rds') to the Nonmem simulation results. If thesepaths have changed, or for other reasons this doesn't work,you can use the 'dir.sims' argument to specify where to findthe Nonmem simulation results. If an '.fst' file was alreadygenerated and is found next to the '_MetaData.rds', the pathto the Nonmem simulation results is not used.

If simulations seem to not be done yet, wait for themto finish? If not, an error will be thrown. If you choose towait, the risk is results never come. 'NMreadSim' will bewaiting for an 'lst' file. If Nonmem fails, it will normallygenerate an 'lst' file. But if 'NMTRAN' fails (checks ofcontrol stream prior to running Nonmem), the 'lst' file is notgenerated. Default is not to wait.

Skip models where results are not available?Default is 'FALSE' meaning an error will be thrown if one ormore models do not have completed results.

Turn off some messages about what is going on?Default is to report the messages.

Track progress? Default is 'TRUE' if 'quiet' isFALSE and more than one model is being simulated. The progresstracking is based on the number of models completed, not thestatus of the individual models.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

If found, check whether 'fst' file modificationtime is newer than 'rds' file. The 'fst' is generated based oninformation in ‘rds', but notice that some systems don’tpreserve the file modification times. Becasue of that,'check.time' is 'FALSE' by default.

By default, 'NMreadSim' will use information aboutthe relative path from the results table file('_MetaData.rds') to the Nonmem simulation results. If thesepaths have changed, or for other reasons this doesn't work,you can use the 'dir.sims' argument to specify where to findthe Nonmem simulation results. If an '.fst' file was alreadygenerated and is found next to the '_MetaData.rds', the pathto the Nonmem simulation results is not used.

If simulations seem to not be done yet, wait for themto finish? If not, an error will be thrown. If you choose towait, the risk is results never come. 'NMreadSim' will bewaiting for an 'lst' file. If Nonmem fails, it will normallygenerate an 'lst' file. But if 'NMTRAN' fails (checks ofcontrol stream prior to running Nonmem), the 'lst' file is notgenerated. Default is not to wait.

Turn off some messages about what is going on?Default is to report the messages.

Skip models where results are not available?Default is 'FALSE' meaning an error will be thrown if one ormore models do not have completed results.

Track progress? Default is 'TRUE' if 'quiet' isFALSE and more than one model is being read. The progresstracking is based on the number of models completed/read, notthe status of the individual models.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

a data set or a fst file

A path to an fst file or a data set

A data.frame with new initial estimates, same styleas returned by NMdata::NMreadExt. Column' par.type' can containelements THETA, OMEGA, SIGMA.

Passed to NMdata::NMwriteSection. This is important forNMreplaceInits to run at all.

Just the file name, not including path

A data.frame containing model paths etc as createdbyNMsim().

Number of seeds in each simulation controlstream. Default is to match length of dist.

Distribution of random sources. These characterstrings will be pasted directly into the Nonem control streamsafter the seed values. Default is "" which means one normaldistribution.dist=c("","UNIFORM") will give two seedswith random sources following a normal and a uniformdistribution.

Optionally, seed values. This can be a data.framewith as many columns as random sources.

Path(s) to the input control stream(s) to run thesimulation on. The output control stream is for now assumed tobe stored next to the input control stream and ending in .lstinstead of .mod. The .ext file must also be present. Ifsimulating known subjects, the .phi is necessary too.

The simulation data as adata.frame or a listofdata.frames. If a list, the model(s) will be run oneach of the data sets in the list.

Number of subproblems to use asSUBPROBLEMS in$SIMULATION block in Nonmem. Thedefault is subproblem=0 which means not to useSUBPROBLEMS.

If simulation results found on file, shouldthey be used? If TRUE and reading the results fail, thesimulations will still be rerun.

A value passed toset.seed(). It isrecommended to useseed.R rather than callingset.seed() manually because the seed can then becaptured and stored byNMsim() for reproducibility. Seeseed.nm for finer control of the seeds that are used inthe Nonmem control streams.

Control Nonmem seeds. If a numeric, a vector or a'data.frame', these are used as the the seed values (a singlevalue or vector will be recycled so make sure the dimesnsionsare right, the number of columns in adata.frame willdictate the number of seeds in each Nonmem control stream. Usea list with elements 'values', and 'dist' and others fordetailed control of the random sources. See?NMseed fordetails on what arguments can be passed this way.

Default is to draw seeds betwen0 and 2147483647 (the values supported by Nonmem) for eachsimulation. You can pass a function that will be evaluated(say to choose a different pool of seeds to draw from).

To avoid changing an exisiting seed in a control stream, useseed.nm="asis".

In casemethod.sim=NMsim_EBE, seeds are not used.

Give all filenames related to the simulation asuffix. A short string describing the sim is recommended like"ph3_regimens".

Variables to be printed in output table as acharacter vector or a space-separated string of variablenames. The default is to export the same tables as listed inthe input control stream. Iftable.vars is provided,all output tables in estimation control streams are droppedand replaced by a new one with just the provided variables. Ifmany variables are exported, and much fewer are used, it canspeed up NMsim significantly to only export what is needed(sometimes this is as little as "PRED IPRED"). Nonmem writesdata slowly so reducing output data can make a very bigdifference in execution time. Seetable.options too.

A character vector or a string ofspace-separated options. Only used iftable.vars isprovided. If constructing a new output table withtable.vars the default is to add two options,NOAPPEND andNOPRINT. You can modify that withtable.options. Do not try to modify output filename -NMsim takes care of that. See 'table.format' too.

A format for '$TABLE'. Only used if'table.vars' is provided. Default is "s1PE16.9". NMsim needs ahigh-resolution format. The Nonmem default "s1PE11.4" isinsufficient for simulation data sets of 1e5 rows or more.

Variables from input data that should be includedin results. Default is to include everything. If working withlarge data sets, it may be wanted to provide a subset of thecolumns here. If doing very large simulations, this may alsobe a way to save memory.

A function (not quoted) that creates thesimulation control stream and other necessary files for asimulation based on the estimation control stream, the data,etc. The default is calledNMsim_default which willreplace any estimation and covariance step by a simulationstep. See details section on oter methods, and see examplesand especially vignettes on how to use the different providedmethods.

Run with all ETAs fixed to zero? Technically allETAs=0 is obtained by replacing$OMEGA by a zeromatrix. Default is 'FALSE'. Instead of a logical 'TRUE/FALSE',a character vector can be used to specify what parameter typesto set to zero and fix. Examples:'typical=c("OMEGA","SIGMA")','typical=c("THETAPV","OMEGA","OMEGAP","OMEGAPD")'. In fact, if'typical=TRUE', both '$OMEGA' itself and - if found - theirpriors will be fixed at zero.

Control the parameter values. 'inits' is a list andcontains (any of) the 'method' used to edit the parameters,and what modifications to do.

Using the defaul 'method', all other list elements are passed asarguments to 'NMwriteInits()'. Please see '?NMwriteInits' and theexamples on the NMsim website for how to edit the parametervalues:https://nmautoverse.github.io/NMsim/articles/NMsim-modify-model.html

The 'method' element controls which method is used to do this, andthis corresponds to the old 'method.update.initxfgs'argument. Normally, the user should not need to deal with thisas the default 'nmsim' method is very flexible andpowerful. If using the new 'method=nmsim' you can specifyparameter values, fix/unfix them, and edit lower and upperlimits for estimation.

• 'method="nmsim"' (default) A highly flexible internalmethod, allows for modification of the parameter values. Allother elements in 'inits' are passed to'NMwriteInits()'. Example where 'THETA(2)' is customized:'inits=list("THETA(2)"=list(init=1.3))'. See '?NMwriteInits'too.

• 'method="psn"'Uses PSN's "update_inits". Requires a functioning PSNinstallation and possibly thatdir.psn is correctlyset. The advantages of this method are that it keeps commentsin the control stream and that it is a method known to many.

• 'method="simple"' Uses a simple internal method toupdate the parameter values based on the ext file. Theadvantages are it does not require PSN, and that it does not relyon code-interpretation for generation of simulation controlstreams. "simple" fixes the whole OMEGA and SIGMA matricesas single blocks which is robust because it avoids anyinterpretation of BLOCK structure or other code in the controlstreams. The downside is it strips all comments, and generallymakes the $OMEGA and $SIGMA sections of the simulation controlstreams less easy to read. "simple" can be used as afallback in case of any issues with 'method="nmsim"'.

• 'method="none"' Do nothing. This is useful if the model to simulatehas not been estimated but parameter values have been manually putinto the respective sections in the control stream.

See also 'file.ext' which can now be handled by 'inits' too. Thischange collects the update of the "initial" parameter values intoone interface rather than multiple arguments.

Named list of additional control stream sectionedits. Note, these can be functions that define how to editsections. This is an advanced feature which is not needed torun most simulations. It is however powerful for some types ofanalyses, like modifying parameter values. See vignettes forfurther information.

Edit data filters ('IGNORE'/'ACCEPT' statements)before running model. This should normally only be used if nodata set is provided. It can be useful if simulating for a VPCbut a different subset of data needs to be simulated than theone used for estimation. A common example on this is inclusionof BLQ's in the VPC even if they were excluded in theestimation. See '?NMreadFilters' which returns a table you canedit and pass to 'filters'. You can also just pass a stringrepresenting the full set of filters to be used. If you pass astring, consider including "IGN=@" to avoid character rows,like the column headers.

If needed, adjust the '$SIZES' section by providing alist of arguments to 'NMupdateSizes()'. Example:‘sizes=list(PD=80)'. See '?NMupdateSizes' for details. Don’tuse arguments like 'file.mod' and 'newfile' which are handledinternally.

The path to the Nonmem executable to use. Thecould be something like "/usr/local/NONMEM/run/nmfe75" (whichis a made up example). No default is available. You should beable to figure this out through how you normally executeNonmem, or ask a colleague.

Submit to cluster? Default is not to, but this is veryuseful if creating a large number of simulations,e.g. simulate with all parameter estimates from a bootstrapresult.

Number of cores used in parallelization. Only used if'sge=TRUE'.

Execute the simulation or only prepare it?'execute=FALSE' can be useful if you want to do additionaltweaks or simulate using other parameter estimates.

The path to the script where this is run. Forstamping of dataset so results can be traced back to code.

A list defining transformations to be appliedafter the Nonmem simulations and before plotting. For eachlist element, its name refers to the name of the column totransform, the contents must be the function to apply.

reorder columns by callingNMdata::NMorderColumns before saving dataset andrunning simulations? Default is TRUE.

Specify how to call Nonmem. Options are"psn" (PSN's execute), "nmsim" (an internal method similar toPSN's execute), and "direct" (just run Nonmem directly anddump all the temporary files). "nmsim" has advantages over"psn" that makes it the only supported method whentype.sim="NMsim_EBE". "psn" has the simple advantage that thepath to nonmem does not have to be specified if "execute" isin the system search path. So as long as you know where yourNonmem executable is, "nmsim" is recommended. The default is"nmsim" if path.nonmem is specified, and "psn" if not.

additional options that will be passed tonmfe. It is only used when path.nonmem is available (directlyor using 'NMdataConf()'). Default is "-maxlim=2" For PSN, see'args.psn.execute'.

Include 'NMREP' as counter of subproblems? Thedefault is to do so if 'subproblems>0'. This will insert acounter called 'NMREP' in the '$ERROR' section and includethat in the output table(s). At this point, nothing is done toavoid overwriting existing variables.

Only used if 'data' is provided. Use this if youare including an exclusion flag column in data. However, whatNMsim will then do is to require that column to equal '0'(zero) for the rows to be simulated. It is often better tosubset the data before simulation. See 'filters' too.

The directory in which to find PSN's executables('execute' and 'update_inits'). The default is to rely on thesystem's search path. So if you can run 'execute' and'update_inits' by just typing that in a terminal, you don'tneed to specify this unless you want to explicitly use aspecific installation of PSN on your system.

A charachter string that will be passed asarguments PSN's 'execute'. The default is"-model_dir_name -nm_output=coi,cor,cov,ext,phi,shk,xml -nmfe_options=\"-maxlim=2\""in addition to the "-clean" based on the 'clean'argument. Notice, if 'path.nonmem' is provided, the default isnot to use PSN.

Ifexecute=TRUE&sge=FALSE, NMsimwill normally read the results usingNMreadSim. Usethis argument to pass additional arguments (in a list) to thatfunction if you want the results to be read in a specificway. This can be if the model for some reason drops rows, andyou need to merge by a row identifier. You would do'args.NMscanData=list(col.row="ROW")' to merge by a columncalled 'ROW'. This is only used in rare cases.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

A charachter string, either "windows" or"linux" - case insensitive. Windows is only experimentallysupported. Default is to useSys.info()[["sysname"]].

The directory in which NMsim will store allgenerated files. Default is to create a folder called 'NMsim'next to 'file.mod'.

Provide a path to a directory in which to save rdsfiles with paths to results. Default is to use dir.sims. Afterrunning 'NMreadSim()' on these files, the original simulationfiles can be deleted. Hence, providing both 'dir.sims' and'dir.res' provides a structure that is simple toclean. 'dir.sims' can be purged when 'NMreadSim' has been runand only small 'rds' and 'fst' files will be kept in'dir.res'. Notice, in case multiple models are simulated,multiple 'rds' (to be read with 'NMreadSim()') files will becreated by default. In cases where multiple models aresimulated, see 'file.res' to get just one file refering to allsimulation results.

Path to an rds file that will contain a table ofthe simulated models and other metadata. This is needed forsubsequently retrieving all the results using'NMreadSim()'. The default is to create a file called'NMsim_..._MetaData.rds' under thedir.res directorywhere ... is based on the model name. However, if multiplemodels (file.mod) are simulated, this will result inmultiple rds files. Specifying a path ensures that one rdsfile containing information about all simulated models will becreated. Notice iffile.res is supplied,dir.resis not used.

If 'TRUE' (default) a dedicated subdirectorywill be created for eac model run. This is normally thecleanest way to run simulations. However, when 'NMsim()' isused for estimation, it may be better to provide model resultsin the same folder as the input control stream (like PSN woulddo). Use 'dir.sim.sub=FALSE' to get this behavior.

Wait for simulations to finish? Default is to do so ifsimulations are run locally but not to if they are sent to thecluster. Waiting for them means that the results will be readwhen simulations are done. If not waiting, path(s) to 'rds'files to read will be returned. Pass them through'NMreadSim()'. Conveniently, NMreadSim() also takes the 'wait'argument too, allowing flexibility to run Nonmem in thebackground, and then read the results, still waiting forNonmem to finish.

A character string to be pasted into$SIMULATION. This must not contain seed or SUBPROBLEM which ishandled separately. Default is to include "ONLYSIM". Youcannot avoid that using 'text.sim'. If needed, you can use'onlysim=FALSE' which will be passed to 'NMsim_default()'.

Add a column called 'DV' to input data sets if acolumn of that name is not found? Nonmem is generallydependent on a 'DV' column in input data but this is typicallyuninformative in simulation data sets and hence easilyforgotten when generating simulation data sets. Ifauto.dv=TRUE and no 'DV' column is found, 'DV=NA' willbe added. In this case ('auto.dv=TRUE' and no 'DV' columnfound) a 'MDV=1' column will also be added if none found.

The degree of cleaning (file removal) to do afterNonmem execution. If 'method.execute=="psn"', this is passedto PSN's 'execute'. If 'method.execute=="nmsim"' a similarbehavior is applied, even though not as granular. NMsim'sinternal method only distinguishes between 0 (no cleaning),any integer 1-4 (default, quite a bit of cleaning) and 5(remove temporary dir completely).

If TRUE (default) this will wipe thesimulation directory before running new simulations. Thedirectory that will be emptied is _not_ dir.sims where you maykeep many or all your simulations. It is the subdirectorynamed based on the run name andname.sim. The reason itis advised to wipe this directory is that if you in a previoussimulation created simulation runs that are now obsolete, youcould end up reading those too when collecting theresults. NMsim will delete previously generated simulationcontrol streams with the same name, but this option goesfurther. An example where it is important is if you first ran1000 replications, fixed something and now rand 500. If youchoose FALSE here, you can end up with the results of 500 newand 500 old simulations.

If the directories specified in dir.sims anddir.res do not exists, should it be created? Default is TRUE.

If TRUE, messages from what is going on will besuppressed.

Silent console messages from Nonmem? The defaultbehaviour depends. It is FALSE if there is only one model toexecute and 'progress=FALSE'.

Track progress? Default is 'TRUE' if 'quiet' isFALSE and more than one model is being simulated. The progresstracking is based on the number of models completed, not thestatus of the individual models.

Check the provided control streams for contentsthat may cause issues for simulation. Default is 'TRUE', andit is only recommended to disable this if you are fully awareof such a feature of your control stream, you know how itimpacts simulation, and you want to get rid of warnings.

For development purposes - users donot need this argument. Controls what format the completeinput data set is saved in. Possible values are 'rds'(default), 'fst' (experimental) and 'csv'. 'fst' may be fasterand use less disk space but factor levels may be lost frominput data to output data. 'csv' will also lead to loss ofadditional information such as factor levels.

Deprecated. Use 'table.vars' and 'table.options'instead.

Deprecated. Use name.sim instead.

Deprecated. Seeseed.R andseed.nm.

Deprecated. Use'inits=list(file.ext="path/to/file.ext")' instead. Optionallyprovide a parameter estimate file from Nonmem. This isnormally not needed since 'NMsim' will by default use the extfile stored next to the input control stream (replacing thefile name extension with '.ext'). If usingmethod.update.inits="psn", this argument cannot be used.

Deprecated, please migrate to 'inits'instead. The initial values of all parameters are by updatedfrom the estimated model before running the simulation. NMsimcan do this with a native function or use PSN to do it - orthe step can be skipped to not update the values.

Deprecated. Use modify instead.

Deprecated. Use modify instead.

Additional arguments passed tomethod.sim.

A control stream to check

The control stream as text lines. Only use of of'file.mod' and 'lines'.

See ?NMsim

See ?NMsim

See ?NMsim

Throw an error if the configuration does not seemto match system.

See ?NMsim

The path to the control stream to be edited. This function overwrites the contents of the file pointed to by file.sim.

Path to the path to the original input control stream provided as 'file.mod' to 'NMsim()'.

See?NMsim.

A phi file to take the known subjects from. Thedefault is to replace the filename extension on file.mod with.phi. A different .phi file would be used if you want to reusesubjects simulated in a previous simulation.

If TRUE, just the text will be returned, andresulting control stream is not written to file.

The path to the control stream to be edited. Thisfunction overwrites the contents of the file pointed to byfile.sim.

Path to the path to the original input controlstream provided as 'file.mod' to 'NMsim()'.

Included for compatibility with 'NMsim()'. Notused.

Used in$PRIOR NWPRI PLEV=0.999. This is aNONMEM argument to the NWPRI subroutine. When PLEV < 1, avalue of THETA will actually be obtained using a truncatedmultivariate normal distribution, i.e. from an ellipsoidalregion R1 over which only a fraction of mass of the normaloccurs. This fraction is given by PLEV.

A umeric value to add to the diagonal of thecovariance matrix. This can be used in case of negativeeigenvaluen in variance-covariance matrix.

Additional arguments passed to 'NMsim_default()'.

The path to the control stream to be edited. Thisfunction overwrites the contents of the file pointed to byfile.sim.

Path to the path to the original input controlstream provided as 'file.mod' to 'NMsim()'.

Included for compatibility with 'NMsim()'. Notused.

Number of replications wanted. The default is 1. Ifgreater, multiple control streams will be generated.

When 'ext' is not used, parameters aresampled, using 'samplePars()'. 'method' must be either'mvrnorm' or 'simpar'. Only used when 'ext' is not provided.

Parameter values in long format as created by'readParsWide' and 'NMdata::NMreadExt'.

If supplied, a path to an rds file where theparameter values used for simulation will be saved.

Additional arguments passed to 'NMsim_default()'.

See?NMsim.

See?NMsim.

See?NMsim.

See?NMsim.

See?NMsim.

See?NMsim.

Number of replications wanted. The default is 1. Ifgreater, multiple control streams will be generated.

Include 'ONLYSIM' in '$SIMULATION'? Default is'TRUE'. Only applied when 'replace.sim='TRUE'.

If there is a $SIMULATION section in thecontents of file.sim, should it be replaced? Default isyes. See thelist.section argument toNMsim forhow to provide custom contents to sections withNMsiminstead of editing the control streams beforehand.

If TRUE, just the text will be returned, andresulting control stream is not written to file.

Everything passed to NMsim_EBE()

See?NMsim.

See?NMsim.

See?NMsim.

If TRUE, just the text will be returned, andresulting control stream is not written to file.

a control stream, path or 'NMctl' object.

What section to update

Model name

The file name extension of the file name to beupdated (e.g., one of "tab", "csv", "msf").

Addditional text to insert right after$SECTION. It can be additional TABLE variables.

The Nonmem parameter that specifies the file. In$TABLE, this is FILE. In $EST it's probably MSFO.

This is used to overwrite the contents of the section. The section output file name will still handled/updated.

Suppress messages? Default is 'FALSE'.

The control stream to update. Will not be edited.

Path to ext file. Default is to replace extension on 'file.mod'.

New file to generate

Path to control stream. Use 'file' or 'lines'.

Control stream as text lines. Use 'file' or 'lines'.

A data frome with filters, like returned by'NMreadFilters()'.

If 'file' is provided, write the results to file? If'lines' is used, 'write' cannot be used.

Path to control stream.

Control stream as character vector. Use either'file.mod' or 'lines', not both.

If 'TRUE' (default), the parameter values areupdated based on the '.ext' file. The path to the '.ext' filecan be specified with 'file.ext' but that is normally notnecessary.

Optionally provide the path to an '.ext' file. Ifnot provided, the default is to replace the file nameextention on 'file.mod' with '.ext'. This is only used if'update=TRUE'.

An long-format parameter table as returned by'NMreadExt()'. Can contain multiple models if 'file.mod' doesnot.

A wide-format parameter table, well suited forcustomizing initial values, limits, and for fixingparameters. For multiple custom parameter specifications, thismay be the most suitable argument.

A list of lists. Each list specifies a parameterwith named elements. Must be named by the parametername. 'lower', 'upper' and 'fix' can be supplied to modify theparameter. See examples. Notice, you can use '...'instead. 'values' may be easier for programming but other thanthat, most users will find '...' more intuitive.

If provided, the results are written to this fileas a new input control stream.

Parameter specifications. See examples,

A path to a control stream. See also alternative'lines' argument. Notice, if 'write' is 'TRUE' (default) and'newfile' is not provided, 'file.mod' will be overwritten.

An optional path to write the resulting controlstream to. If nothing is provided, the default is to overwrite'file.mod'.

Control stream lines as a character vector. If youalready read the control stream - say using'NMdata::NMreadSection()', use this to modify the text lines.

The default behavior ('wipe=FALSE') is to add the'$SIZES' values to any existing values found. If SIZESparameter names are overlapping with existing, the values willbe updated. If 'wipe=TRUE', any existing '$SIZES' section isdisregarded.

Write results to 'newfile'?

The $SIZES parameters. Provided anything, like 'PD=40'See examples.

Parameters in long format, as returned by'NMreadExt()'.

Elements to add.

Either "top" or "bottom". Decides if new text isprepended or appended to existing text.

The object to add class to

The class to add (character)

Nonmem-style data set. If using 'TAPD' an 'EVID'column must contain 1 for dosing records.

A numerical vector with simulation times. Can also bea data.frame in which case it must contain a 'TIME' column andis merged with 'data'.

A numerical vector with simulation times, relative toprevious dose. When this is used, 'data' must contain rowswith 'EVID=1' events and a 'TIME' column. 'TAPD' can also be adata.frame in which case it must contain a 'TAPD' column andis merged with 'data'.

The compartment in which to insert the EVID=2records. Required if 'CMT' is a column in 'data'. If longerthan one, the records will be repeated in all the specifiedcompartments. If a data.frame, covariates can be specified.

The value to put in the 'EVID' column for the createdrows. Default is 2 but 0 may be prefered even for simulation.

Optionally provide a single value to be assigned to the'DV' column. The default is to assign nothing which willresult in 'NA' as samples are stacked ('rbind') with'data'. If you assign a different value in 'DV', the defaultvalue of 'EVID' changes to '0', and 'MDV' will be '0' insteadof '1'. An example where this is useful is when generatingdatasets for '$DESIGN' where 'DV=0' is often used.

The name of the column in 'data' that holds theunique subject identifier.

Only relevant - and likely not needed -if data contains ADDL and II columns. If those columns areincluded, 'addEVID2()' will use 'NMdata::NMexpanDoses()' toevaluate the time of each dose. Other than the 'data'argument, 'addEVID2()' relies on the default 'NMexpanDoses()'argument values. If this is insufficient, you can specifyother argument values in a list, or you can call'NMdata::NMexpanDoses()' manually before calling 'addEVID2()'.

If 'TRUE' (default), events are reduced to uniquetime points before insertion. Sometimes, it's easier tocombine sequences of time points that overlap (maybe across'TIME' and 'TAPD'), and let 'addEVID2()' clean them. If youwant to keep your duplicated events, use 'unique=FALSE'.

If 'TIME' and/or 'TAPD' are 'data.frame'sand contain other columns than 'TIME' and/or 'TAPD', those areby default assumed to be covariates to be merged withdata. More specifically, they will be merged by when thesample times are added. If 'extras.are.covs=FALSE', they willnot be merged by. Instead, they will just be kept asadditional columns with specified values, aligned with thesample times.

The default is to return data as a'data.frame'. Pass a function (say 'tibble::as_tibble') inas.fun to convert to something else. If data.tables arewanted, use 'as.fun="data.table"'. The default can beconfigured using 'NMdataConf()'.

Deprecated. Use 'data'.

Deprecated. Use 'TIME'.

Table of parameters to augment with additional columns

Optional string to add to all new columnnames. Maybe except 'i' and 'j'.

Add 'i' and 'j'? Default is 'TRUE' if no suffix is supplied, and 'FALSE' if a suffix is specified.

Overwrite non-missing values? Default is 'FALSE'.

Table of parameters to augment with additional columns.

Overwrite non-missing values? Default is 'FALSE'.

A data set containing indiviudual predictions. Often aresult of NMsim.

Path to the ext file to take the parameterestimates from.

Parameter number of parameter holding variance of theproportional error component. If ERR(1) is used forproportional error, use prop=1. Can also refer to a thetanumber.

Parameter number of parameter holding variance of theadditive error component. If ERR(1) is used for additiveerror, use add=1. Can also refer to a theta number.

Should the error be added on log scale? This is used toobtain an exponential error distribution.

Use "sigma" if variances are estimated with theSIGMA matrix. Use "theta" if THETA parameters are used. See'scale.par' too.

If log=FALSE, truncate simulated values at 0? Iftrunc0, returned predictions can be negative.

Denotes if parmeter represents a variance or astandard deviation. Allowed values and default value dependson 'par.type'.

• if par.type="sigma" only "var"is allowed.

• if par.type="theta" allowed values are "sd"and "var". Default is "sd".

A character string with an expression denoting asubset in which to add the residual error. Example:subset="DVID=='A'"

A number to pass to set.seed() beforesimulating. Default is to generate a seed and report it in theconsole. Use seed=FALSE to avoid setting the seed (if youprefer doing it otherwise).

The name of the column containing individualpredictions.

The name of the column to be created byaddResVar to contain the simulated observations (individualpredictions plus residual error).

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

Table of initial values as created by NMreadInits().

object to convert

Not used

Path to Nonmem-created file. Typically an outputcontrol stream.

Scan input data for updates too? Default is TRUE.

If you do want to take into account input databut avoid re-reading the information, you can pass the NMdatameta data object.

The input control stream

If files are moved around on or between filesystems, the file modification time may not be reflective ofthe Nonmem runtime. In that case, you can choose to extractthe time stamp from the output control stream. The issue isthat Nonmem does not write the time zone, so you have to passthat to checkTimes if this is wanted.

a string to clean

A covariate specififed in a list. See?expandCovLists.

See ?expandCovLists.

The subject ID column name. Necessary becausequantiles sould be quantiles of distribution of covariate onsubjects, not on observations (each subject contributes once).

Used for rounding of covariate values if usingquantiles or if using a function to find reference.

elements as detected by 'NMreadInits()'

The directory in which to look for contents to clean

The sources to delete temporary content from. Thisis a character vector, and the defailt is'c("nmsim","psn","psnfit","backup")'. Each of these correspondto a preconfigured pattern.

Look recursively in folder? Notice, matches willbe deleted recursively (they are oftendirectories). 'recursive' controls whether they are searchedfor recursively.

Delete the found matches? If not, the matches arejust reported, but nothing deleted.

Passed to 'forestDefineCovs()'

list of (input or output) control streams to consider

The file name or file names to modify.

A character string or a numeric to add to the filename. If a vector, the vector is collapsed to a single string,using 'sep' as separator in the collapsed string.

In case x is numeric, a number of zeros to pad beforethe appended number. This is useful if you are generating saymore than 10 files, and your counter will be 01, 02,..,10,... and not 1, 2,...,10,...

The separator between the existing file name (untilextension) and the addition.

If 'x' is of length greater than 1, the default isto collapse the elements to a single string using 'sep' asseparator. See the 'collapse' argument to '?paste'. If youwant to treat them as separate strings, use 'collapse=NULL'which will lead to generation of separate file names. However,currently 'fn' or 'x' must be of length 1.

"append" (default) or "prepend".

Allow 'fn' to be string(s) without extensions?Default is 'FALSE' in which case an error will be thrown if'fn' contains strings without extensions. If 'TRUE', 'x' willbe appended to fn in these cases.

Covariates provided as lists - see examples. The nameof the arguement must match columns in data set. An elementcalled ref must contain either a reference value or a functionto use to derive the reference value from data(e.g. 'median'). Provide either 'values' or 'quantiles' todefine the covariate values of interest (typically, the valuesthat should later be simulated and maybe shown in a forestplot). 'label' is optional - if missing, the argument namewill be used. If quantiles are requested, they are derivedafter requiring unique values for each subject.

A data set needed if the reference(s) value of one ormore covariates is/are provided as functions (like median), orif covariate values are provided as quantiles.

The subject ID column name. Necessary becausequantiles sould be quantiles of distribution of covariate onsubjects, not on observations (each subject contributes once).

Used for rounding of covariate values if usingquantiles or if using a function to find reference.

If 'TRUE' (default), only return one row withall reference values. If 'FALSE' there will be one such rowfor each covariate. When reduced to one line, all columnsrelated to covariate-level information such as covariate namewill contain 'NA' for the reference.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

Simulated data to process. This data.frame mustcontain must contain multiple columns, as defined by'NMsim::forestDefineCovs()'.

A named list of functions to apply forderivation of exposure metrics.

The coverage of the confidence intervals. Defaultis 0.95.

a character vector of column names to perform allcalculations by. This could be sampling subsets or analyte.

The default is to return data as a'data.frame'. Pass a function (say 'tibble::as_tibble') inas.fun to convert to something else. If data.tables arewanted, use 'as.fun="data.table"'. The default can beconfigured using 'NMdataConf()'.

A dataset that contains "ID" and all 'ETA's. This canbe obtained by 'NMdata::NMscanData'.

Path to the .phi file to be written.

If 'file' exists already, overwrite it? Defaultis 'FALSE'.

The elements object produced by 'NMreadInits()'.

Path to input or output control stream.

NMsim results (class 'NMsimRes').

A list

a data.table

List of coontrol streams as lines

File name to provide stem for all file names

List of objects to provide names for

If only one file path, skip numbering? Default isTRUE.

Passed to 'gsub()'

This is passed to gsub(), but ‘overwrite()'’s defaultbehavior is the opposite of the one of 'gsub()'. Default is'FALSE' which means that strings that are exactly matched willbe replaced. This is useful because strings like 'THETA(1)'contains special characters. Use 'fixed=FALSE' to use regularexpressions. Also, see other arguments accepted by 'gsub()'for advanced features.

integers to pad. They can be coded as characters already.

Optional specification of length of characterstrings to return. If not supplied, characters will be paddedto match length of max value in x.

A vector of strings

A string to add

Aditional arguments to 'paste()'.

Output of NMsim::NMcreateMatLines. This isa string of OMEGA/SIGMA estimates that will be wrapped ontomultiple lines for ease of reading in NONMEM control streams.

The summary object to be printed. See ?summary.NMsimRes

Arguments passed to other print methods.

vector of file paths. Typically to Nonmemexecutables.

If TRUE, an error is thrown if no paths arevalid.

object to read.

Not used.

A data.frame or a path to a delimited file to be readusing 'data.table::fread'.

Column containing name of the original model. Bydefault a column called "model" will contain "Model1".

Name of the model counter, default is"model.sim". If the provided name is not found in data, itwill be created as a row counter. Why needed? Each row in datarepresents a set of parameters, i.e. a model. In the longformat result, each model will have multiple rows. Hence, amodel identifier is needed to distinguish between models inresults.

Defines how column names get associatedwith THETA, OMEGA, and SIGMA. Default is to look for "T", "O",or "S" as starting letter. If customizing, make sure each nocolumn name will be matched by more than one criterion.

The default is to return data as a data.frame. Passa function (saytibble::as_tibble) in as.fun to convertto something else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

A simulation data set with only one subject

The number of subjects to be sampled. This can begreater than the number of subjects in data.covs.

Name of the subject ID column in 'data' (default is"ID").

Name of the subject ID column in 'data.covs'(default is "ID").

The data set containing the subjects to samplecovariates from.

The name of the covariates (columns) to sample from'data.covs'.

If provided, passed to 'set.seed()'.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

Path to model control stream. Will be used forboth 'NMreadExt()' and 'NMreadCov()', and extension willautomatically be replaced by '.ext' and '.cov'.

Number of sets of parameter values togenerate. Passed to 'simpar'.

The sampling method. Options are "mvrnorm" and"simpar". Each have pros and cons. Notice that both methodsare fully automated as long as ".ext" and ".cov" files areavailable from model estimation.

seed value passed to set.seed().

The returned data set format "ext" (default) or"wide". "ext" is a long-format, similar to what'NMdata::NMreadExt()' returns.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

Path to model control stream. Will be used forboth 'NMreadExt()' and 'NMreadCov()', and extension willautomatically be replaced by '.ext' and '.cov'.

Number of sets of parameter values to generate. Passedto 'simpar'.

"ext" (default) or "wide".

seed value passed to set.seed().

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

Passed to 'NMdata::NMreadExt()'. Path to ext file. Bydefault, 'NMreadExt()' uses a'auto.ext=TRUE' which means thatthe file name extension is replaced by '.ext'. If your extfile name extension is not '.ext', add 'auto.ext=FALSE' (see...).

Number of subjects to generate

Optional seed. Will be passed to 'set.seed'. Samething as running 'set.seed' just before calling'simPopEtas()'.

A long-format parameter table containing par.type andi columns. If this is supplied, the parameter values will notbe read from an ext file, and file has no effect. If an extfile is available, it is most likely better to use the fileargument.

An optional phi file to write the generatedsubjects to.

If 'file.phi' exists already, overwrite it?Default is 'FALSE'.

The default is to return data as a data.frame. Passa function (say 'tibble::as_tibble') in as.fun to convert tosomething else. If data.tables are wanted, useas.fun="data.table". The default can be configured usingNMdataConf.

Deprecated. Use file instead.

Deprecated. Use seed.R instead.

Additional arguments passed to NMdata::NMreadExt().

Name of the argument

argument value

If val.arg is NULL, what should be returned?

What values are allowed

run tolower?

clean white spaces?

single or multiple file or dir paths as strings.

Passed to 'forestSummarize()'

An NMsimRes object (from NMsim).

Not used

The length of the diagonal. Same as number of rowsor columns.

path to control stream to edit

control stream as lines. Use either file.sim orlines.sim.

The sections (parameter types) to edit. Default is'c("OMEGA", "OMEGAP", "OMEGAPD")'.

path and filename to new run. If missing or NULL,output is returned as a character vector rather than written.

an NMsimModTab object

arguments passed to other methods.

an NMsimRes object

arguments passed to other methods.

the character lines to write

The file name path to write to

Passed to 'nameMultipleFiles()'