find.value.fun: Find the last occurence of the higher/lower value in a...

In Dataset: Handling survey data in R. (official release)

Description

The find.value.fun function allows to find the last occurence of the higher/lower value in a set of columns of a data.frame object, and returns these values. You can then retrieve these last values for other sets of columns of the data.frame object.

Usage

find.value.fun(data, from, to, fun = "max")

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 occurence of the 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.
fun	a character, the name of the operator to use: currently 'max' or 'min'.

Details

This function is useful when working on longitudinal data with attrition problems. It allows to retrieve the last occurence of the higher/lower value in a specific time-varying variable and then getting the values for other time-varying variables according to the indexs previously reported. For numerical time-varying variable, the numeric order is used. For factors, order of levels has to be specified in the factor corresponding to the first year (see example below).

Value

If the to argument is missing, the function returns the ids where the last occurence of the higher/lower value is for each row of the data argument. It also return a column containing the value retrieved. NA will be returned for both columns if the sequence contains only NA. If the to argument is given, a data.frame containing the corresponding last values on covariates is returned.

Author(s)

Emmanuel Rousseaux

See Also

findlast, find.value.last, seq.count.

Examples

fv.data.example <- data.frame(
  'unemployment99' = c(0,NA,NA,NA,NA),
  'unemployment00' = c(0,NA,NA,NA,NA),
  'unemployment01' = c(1,0,NA,NA,NA),
  'unemployment02' = c(0,0,0,NA,NA),
  'unemployment03' = c(0,1,NA,NA,NA),
  'unemployment04' = c(NA,1,1,NA,NA),
  'unemployment05' = c(NA,1,1,NA,NA),
  'unemployment06' = c(NA,NA,1,NA,NA),
  'unemployment07' = c(NA,NA,NA,0,NA),
  'unemployment08' = c(NA,NA,NA,0,NA),
  'unemployment09' = c(NA,NA,NA,0,NA),
  'unemployment10' = c(NA,NA,NA,0,NA),
  'unemployment11' = c(NA,NA,NA,0,NA),
  'age99' = c(15,NA,NA,NA,NA),
  'age00' = c(16,NA,NA,NA,NA),
  'age01' = c(17,20,NA,NA,NA),
  'age02' = c(18,21,23,NA,NA),
  'age03' = c(19,22,NA,NA,NA),
  'age04' = c(NA,23,25,NA,NA),
  'age05' = c(NA,24,26,NA,NA),
  'age06' = c(NA,NA,27,NA,NA),
  'age07' = c(NA,NA,NA,29,NA),
  'age08' = c(NA,NA,NA,30,NA),
  'age09' = c(NA,NA,NA,31,NA),
  'age10' = c(NA,NA,NA,32,NA),
  'age11' = c(NA,NA,NA,33,NA),
  'education99' = c(8,NA,NA,NA,NA),
  'education00' = c(8,NA,NA,NA,NA),
  'education01' = c(8,3,NA,NA,NA),
  'education02' = c(8,3,2,NA,NA),
  'education03' = c(8,4,NA,NA,NA),
  'education04' = c(NA,4,2,NA,NA),
  'education05' = c(NA,4,2,NA,NA),
  'education06' = c(NA,NA,3,NA,NA),
  'education07' = c(NA,NA,NA,7,NA),
  'education08' = c(NA,NA,NA,7,NA),
  'education09' = c(NA,NA,NA,7,NA),
  'education10' = c(NA,NA,NA,8,NA),
  'education11' = c(NA,NA,NA,8,NA),
  'health99' = factor(c('very well',NA,NA,NA,NA), levels = c("poor", "well", "very well")), # we specify the order of levels to use in the first year of the variable
  'health00' = c('very well',NA,NA,NA,NA),
  'health01' = c('very well','well',NA,NA,NA),
  'health02' = c('very well','well','poor',NA,NA),
  'health03' = c('well','very well',NA,NA,NA),
  'health04' = c(NA,'very well','poor',NA,NA),
  'health05' = c(NA,'very well','poor',NA,NA),
  'health06' = c(NA,NA,'poor',NA,NA),
  'health07' = c(NA,NA,NA,'very well',NA),
  'health08' = c(NA,NA,NA,'very well',NA),
  'health09' = c(NA,NA,NA,'poor',NA),
  'health10' = c(NA,NA,NA,'poor',NA),
  'health11' = c(NA,NA,NA,'well',NA)
)

# we get column ids where the last occurence of the higher value is
find.value.fun(
  data = fv.data.example,
  from = 1:13 # for unemployment
)
find.value.fun(
  data = fv.data.example,
  from = 40:52 # for health
)

# we retrieve the last occurence of the higher value for unemployment, and get corresponding values for the variables age, education and health.
find.value.fun(
  data = fv.data.example,
  from = 1:13,
  to = list(
    'age' = 14:26,
    'education' = 27:39,
    'health' = 40:52
  )
)

# we retrieve the last occurence of the higher value for health, and get corresponding values for the variables unemployment, age and education.
find.value.fun(
  data = fv.data.example,
  from = 40:52,
  to = list(
    'age' = 14:26,
    'education' = 27:39,
    'health' = 40:52
  )
)

# same operation but here for the min
find.value.fun(
  data = fv.data.example,
  from = 40:52,
  to = list(
    'age' = 14:26,
    'education' = 27:39,
    'health' = 40:52
  ),
  fun = 'min'
)

Dataset documentation built on May 2, 2019, 6:09 p.m.