miceadds: subset_datlist – R documentation

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

subset_datlist

Subsetting Multiply Imputed Datasets and Nested Multiply Imputed Datasets

Description

Returns a subsets of multiply imputed datasets or nested multiply imputed datasets. These function allows choosing parts of the imputed datasets using the index argument for multiply imputed datasets and index_between and index_within for nested multiply imputed datasets as well as the application of the base::subset S3 method for selecting cases and variables in datasets.

Usage

subset_datlist(datlist, subset=TRUE, select=NULL, expr_subset=NULL,
        index=NULL, toclass="datlist")

## S3 method for class 'datlist'
subset(x, subset, select=NULL, expr_subset=NULL,
                     index=NULL, ...)
## S3 method for class 'imputationList'
subset(x, subset, select=NULL, expr_subset=NULL,
                     index=NULL, ...)
## S3 method for class 'mids'
subset(x, subset, select=NULL, expr_subset=NULL,
                     index=NULL, ...)
## S3 method for class 'mids.1chain'
subset(x, subset, select=NULL, expr_subset=NULL,
                     index=NULL, ...)

subset_nested.datlist( datlist, subset=TRUE, select=NULL, expr_subset=NULL,
      index_between=NULL, index_within=NULL, toclass="nested.datlist",
          simplify=FALSE )

## S3 method for class 'nested.datlist'
subset(x, subset, select=NULL, expr_subset=NULL,
                index_between=NULL, index_within=NULL, simplify=FALSE, ...)
## S3 method for class 'NestedImputationList'
subset(x, subset, select=NULL, expr_subset=NULL,
                index_between=NULL, index_within=NULL, simplify=FALSE, ...)

Arguments

`datlist`	For `subset_datlist` it is a list of datasets or an object of class `datlist`, `imputationList`, `mids` or `mids.1chain`. For `subset_nested.datlist` it is a list of datasets or an object of class `nested.datlist` or `NestedImputationList`.
`subset`	Logical expression indicating elements or rows to keep, see `base::subset`. `subset` can also be a numeric vector containing row indices.
`select`	Expression indicating columns to select from a data frame
`expr_subset`	Expression indicating a selection criterion for selection rows.
`index`	Vector of indices indicating which of the multiply imputed datasets should be selected.
`toclass`	The object class in which the datasets should be saved.
`index_between`	Index for between nest datasets
`index_within`	Index for within nest datasets
`simplify`	Optional logical indicating whether a nested multiply imputed dataset should be simplified to a multiplied imputed dataset.
`x`	Object containing multiply imputed or nested multiply imputed datasets
`...`	Further arguments to be passed.

Value

For multiply imputed datasets: Object of class datlist, imputationList or mids
For nested multiply imputed datasets: Object of class nested.datlist or NestedImputationList.

Note

If subsetting is applied to objects of class mids (or mids.1chain), then informations about the imputation procedure are lost.

Examples

#############################################################################
# EXAMPLE 1: Subsetting and selection of multiply imputed datasets
#############################################################################

data(data.ma02)

# define original list of datasets
datlist1a <- data.ma02
# object of class datlist
datlist1b <- miceadds::datlist_create(datlist1a)
datlist1b
# object of class imputationList
datlist1c <- mitools::imputationList(datlist1a)
datlist1c
# object of class mids
datlist1d <- miceadds::datlist2mids(datlist1a)
datlist1d

# select some imputed datasets
datlist2a <- miceadds::subset_datlist( datlist1a, index=c(5,3,7) )
datlist2a
# convert to class imputationList
datlist2b <- miceadds::subset_datlist( datlist1a, index=c(5,3,7),
                      toclass="imputationList")
datlist2b
# convert to class mids
datlist2c <- miceadds::subset_datlist( datlist1a, index=1:3, toclass="mids")
datlist2c

## Not run: 
# select some variables
datlist3a <- miceadds::subset_datlist( datlist1a, select=c("idstud", "books")  )
datlist3a
# Because datlist1b is a datlist it is equivalent to
datlist3b <- subset( datlist1b, select=c("idstud", "books")  )
datlist3b
# operating on imputationList class
datlist3c <- miceadds::subset_datlist( datlist1c, select=c("idstud", "books")  )
datlist3c
# operating on mids class
datlist3d <- miceadds::subset_datlist( datlist1d, select=c("idstud", "books")  )
datlist3d
# selection of rows and columns in multiply imputed datasets
datlist4a <- miceadds::subset_datlist( datlist1a, index=1:5,
                  subset=datlist1a[[1]]$idschool < 1067,
                  select=c("idstud", "idschool","hisei") )
datlist4a
# convert to class mids
datlist4b <- miceadds::subset_datlist( datlist1a, index=1:5,
                  subset=datlist1a[[1]]$idschool < 1067,
                  select=c("idstud", "idschool","hisei"), toclass="mids" )
datlist4b
# The same functionality, but now applying to object of class mids datlist1d
datlist4c <- miceadds::subset_datlist( datlist1d, index=1:5,
               subset=datlist1a[[1]]$idschool < 1067,
               select=c("idstud", "idschool","hisei") )
datlist4c

# expression for selecting rows specific in each data frame
# which can result in differently sized datasets (because the variable
# migrant is imputed)
datlist5a <- miceadds::subset_datlist( datlist1a,  expr_subset=expression(migrant==1) )
datlist5a

# select the first 100 cases
datlist6a <- miceadds::subset_datlist( datlist1a, select=c("idstud", "books"),
                       subset=1:100 )
datlist6a

#############################################################################
# EXAMPLE 2: Subsetting and selection of nested multiply imputed datasets
#############################################################################

library(BIFIEsurvey)
data(data.timss4, package="BIFIEsurvey")
dat <- data.timss4

# create object of class 'nested.datlist'
datlist1a <- miceadds::nested.datlist_create( dat )
# create object of class 'NestedImputationList'
datlist1b <- miceadds::NestedImputationList(dat)

# select some between datasets
datlist2a <- subset_nested.datlist( datlist1a, index_between=c(1,3,4) )
datlist2a
# shorter version
datlist2b <- subset( datlist1a, index_between=c(1,3,4) )
datlist2b
# conversion of a NestedImputationList
datlist2c <- subset( datlist1b, index_between=c(1,3,4))
datlist2c
# select rows and columns
sel_cases <- datlist1a[[1]][[1]]$JKZONE <=42
datlist3a <- subset( datlist1a, subset=sel_cases,
                 select=c("IDSTUD","books", "ASMMAT") )
datlist3a
# remove within nest
datlist4a <- subset( datlist1a, index_within=1 )
datlist4a
# remove within nest and simplify structure
datlist4b <- subset( datlist1a, index_within=1, simplify=TRUE)
datlist4b
datlist4c <- subset( datlist1b, index_within=1, simplify=TRUE)
datlist4c
# remove between nest
datlist5a <- subset( datlist1a, index_between=1, simplify=TRUE)
datlist5a
datlist5b <- subset( datlist1b, index_between=1, simplify=TRUE)
datlist5b

## End(Not run)

miceadds

Some Additional Multiple Imputation Functions, Especially for 'mice'

v3.11-6

GPL (>= 2)

Authors

Alexander Robitzsch [aut,cre] (<https://orcid.org/0000-0002-8226-3132>), Simon Grund [aut] (<https://orcid.org/0000-0002-1290-8986>), Thorsten Henke [ctb]

Initial release

2021-01-21 11:48:47