SSBtools (version 1.8.9)

Extend0_n_rnd_groups: Add zero-frequency rows using complete and sampled group combinations

Description

Extend0_with_n_rnd_groups is a function that calls Extend0() with Extend0_n_rnd_groups() as the varGroups attribute. The data are extended in the usual way based on some of the varGroups elements. For the remaining elements, combinations are sampled.

Usage

Extend0_n_rnd_groups(
  data,
  varGroups,
  n_rnd_groups = 0,
  rnd_rep = 1,
  rndSeed = 123
)
Extend0_with_n_rnd_groups(data, varGroups, non_rnd, rnd_rep = 1, rndSeed = 123)

Value

A data frame.

Arguments

data

data.frame within Extend0

varGroups

argument to Extend0

n_rnd_groups

The last n_rnd_groups elements of varGroups are used as the basis for sampling.

rnd_rep

When rnd_rep = 1 (the default), the usual Extend0 extension is performed for the selected varGroups elements. When rnd_rep > 1, this result is replicated to allow more combinations to be sampled.

rndSeed

Internal random seed to be used

non_rnd

In Extend0_with_n_rnd_groups, the varGroups elements that should not be sampled are specified by name. In this case, varGroups must be a named list.

See Also

Extend0rnd1()

Examples

# Data to be extended 
d <- SSBtoolsData("barcelona2025")[c(3, 6, 12, 18), -5]
rownames(d) <- NULL
d$year <- 2025:2026
d$month <- c("January", rep("August", 3))
d

# varGroups as a named list
varGroups <- list(geo = c("country", "city"), 
                  age = "age", 
                  sex = "sex", 
                  time = c("month", "year"))

a0 <- Extend0(d, varGroups = varGroups)
dim(a0)  # all combinations, 48 rows
 
a1 <- Extend0_with_n_rnd_groups(d, varGroups = varGroups, non_rnd = c("time", "geo"))
a1
dim(unique(a1[c("month", "year", "country", "city")])) # all combinations of time and geo  

unique(a1[c("sex", "month", "year")])        # not all combinations


a2 <- Extend0_with_n_rnd_groups(d, varGroups = varGroups, non_rnd = c("sex", "time", "geo"))
dim(a2)

# all combinations of selected variables 
dim(unique(a2[c("sex", "month", "year", "country", "city")]))
 
# not all combinations of selected variables 
dim(unique(a2[c("age", "month", "year", "country", "city")]))  
 

# effect of rnd_rep
for (rnd_rep in c(1, 2, 3, 5, 10, 50)) 
   print(dim(Extend0_with_n_rnd_groups(d, 
           varGroups = varGroups, non_rnd = c("time", "geo"), rnd_rep = rnd_rep)))