findlast: Find the last valid cases in a 'data.frame' object.

In Rsocialdata0: Handling survey data in R. (beta version, not for use)

Description

The findlast function allows to find the last non-missing value in a set of columns of a data.frame object, and returns these values. You can then retrieve these last values for others sets of columns of the data.frame object.

Usage

findlast(data, from, to)

Arguments

data	a data.frame object.
from	a vector of either column names or column ids from where you want to find where the last non-missing value is.
to	a list containing in each item the column names or ids of another time-varying variable for which you want to get the last value found in data[,from]. The name of each item is used as column name for the output.

Details

This function is useful when working on longitudinal data with attrition problems. It allows to retrieve the last values given by repondants in a specific time-varying variable and then getting the values for other time-varying variables according to the index of last participation of each respondant.

Value

If the to argument is missing, the function returns the ids where the last non-missing value is for each row of the data argument. If the to argument is given, a data.frame containing the corresponding last values on covariates is returned.

Author(s)

Emmanuel Rousseaux

See Also

find.value.last, find.value.fun.

Examples

# data.frame example
fl.data.example <- data.frame(
  'age00' = c(48,NA,34,63,NA,27),
  'age01' = c(NA,27,35,64,NA, NA),
  'age02' = c(50,28,36,NA,NA, NA),
  'sex00' = c('M',NA,'F','M',NA,'M'),
  'sex01' = c(NA,'F','F',NA,NA, NA),
  'sex02' = c('M',NA,'F',NA,'M', NA),
  'nationality00' = factor(c('v1',NA,'v1','v1',NA,'o')),
  'nationality01' = factor(c(NA,'o','v2','o',NA, NA)),
  'nationality02' = factor(c('o',NA,'v2',NA,'o', NA))
)

# we get the column ids where the last non-missing value is
findlast(
  data = fl.data.example,
  from = c('age00', 'age01', 'age02')
)

# we retrieve the last value for age, and get corresponding values for the variables sex and nationality
findlast(
  data = fl.data.example,
  from = c('age00', 'age01', 'age02'),
  to = list(
    'age' = c('age00', 'age01', 'age02'),
    'sex' = c('sex00', 'sex01', 'sex02'),
    'nationality' = c('nationality00','nationality01','nationality02')
  )
)

# same operation but using column ids instead of column names
findlast(
  data = fl.data.example,
  from = 1:3,
  to = list(
    'age' = 1:3,
    'sex' = 4:6,
    'nationality' = 7:9
  )
)

Rsocialdata0 documentation built on May 2, 2019, 5:55 p.m.