from_5_to_3: Function reducing from 5 to 3 categorical variables

In InseeFrLab/rtauargus: Using Tau-Argus from R

Function reducing from 5 to 3 categorical variables

Description

Function reducing from 5 to 3 categorical variables

Usage

from_5_to_3(
  dfs,
  dfs_name,
  totcode,
  hrcfiles = NULL,
  sep_dir = FALSE,
  hrc_dir = "hrc_alt",
  v1 = NULL,
  v2 = NULL,
  v3 = NULL,
  v4 = NULL,
  sep = "_",
  maximize_nb_tabs = FALSE,
  verbose = FALSE
)

Arguments

dfs	data.frame with 5 categorical variables (n >= 3 in the general case)
dfs_name	name of the data.frame in the list provided by the user
totcode	named vector of totals for categorical variables
hrcfiles	named vector indicating the hrc files of hierarchical variables among the categorical variables of dfs
sep_dir	allows forcing the writing of hrc files in a separate folder defaulted to FALSE
hrc_dir	folder where to write the hrc files if forcing the writing in a new folder or if no folder is specified in hrcfiles
v1	allows forcing the value of the first variable to merge when reducing from 5 to 4 dimensions, not specified by default (NULL)
v2	allows forcing the value of the second variable to merge when reducing from 5 to 4 dimensions, not specified by default (NULL)
v3	allows forcing the value of the first variable to merge when reducing from 4 to 3 dimensions, not specified by default (NULL)
v4	allows forcing the value of the second variable to merge when reducing from 4 to 3 dimensions, not specified by default (NULL)
sep	separator used during concatenation of variables
maximize_nb_tabs	specifies whether to prefer selecting hierarchical variables with the most nodes as a priority (TRUE), which generates more tables but of smaller size, or non-hierarchical variables with the least modality (FALSE) to create fewer tables
verbose	prints the different steps of the function to notify the user of the progress, mainly for the general function gen_tabs_5_4_to_3()

Value

a list containing the following components:

• tabs: named list of dataframes with 3 dimensions (n-2 dimensions in the general case) endowed with nested hierarchies

• hrcs5_4: named list of hrc specific to the variable created via the merge when reducing from 5 to 4 dimensions

• hrcs4_3: named list of hrc specific to the variable created via the merge when reducing from 4 to 3 dimensions

• alt_tot5_4: named list of totals when reducing from 5 to 4 dimensions

• alt_tot4_3: named list of totals when reducing from 4 to 3 dimensions

• vars: named list of vectors representing the merged variables during the two steps of dimension reduction

Examples

library(dplyr)
data <- expand.grid(
  ACT = c("Total", "A", "B", "A1", "A2", "B1", "B2"),
  GEO = c("Total", "GA", "GB", "GA1", "GA2", "GB1", "GB2"),
  SEX = c("Total", "F", "M","F1","F2","M1","M2"),
  AGE = c("Total", "AGE1", "AGE2", "AGE11", "AGE12", "AGE21", "AGE22"),
  ECO = c("PIB","Ménages","Entreprises"),
  stringsAsFactors = FALSE,
  KEEP.OUT.ATTRS = FALSE
) %>%
  as.data.frame()

data <- data %>% mutate(VALUE = 1:n())

hrc_act <- "hrc_ACT.hrc"
sdcHierarchies::hier_create(root = "Total", nodes = c("A","B")) %>%
  sdcHierarchies::hier_add(root = "A", nodes = c("A1","A2")) %>%
  sdcHierarchies::hier_convert(as = "argus") %>%
  slice(-1) %>%
  mutate(levels = substring(paste0(level,name),3)) %>%
  select(levels) %>%
  write.table(file = hrc_act, row.names = FALSE, col.names = FALSE, quote = FALSE)

hrc_geo <- "hrc_GEO.hrc"
sdcHierarchies::hier_create(root = "Total", nodes = c("GA","GB")) %>%
  sdcHierarchies::hier_add(root = "GA", nodes = c("GA1","GA2")) %>%
  sdcHierarchies::hier_add(root = "GB", nodes = c("GB1","GB2")) %>%
  sdcHierarchies::hier_convert(as = "argus") %>%
  slice(-1) %>%
  mutate(levels = substring(paste0(level,name),3)) %>%
  select(levels) %>%
  write.table(file = hrc_geo, row.names = FALSE, col.names = FALSE, quote = FALSE)

hrc_sex <- "hrc_SEX.hrc"
sdcHierarchies::hier_create(root = "Total", nodes = c("F","M")) %>%
  sdcHierarchies::hier_add(root = "F", nodes = c("F1","F2")) %>%
  sdcHierarchies::hier_add(root = "M", nodes = c("M1","M2")) %>%
  sdcHierarchies::hier_convert(as = "argus") %>%
  slice(-1) %>%
  mutate(levels = substring(paste0(level,name),3)) %>%
  select(levels) %>%
  write.table(file = hrc_sex, row.names = FALSE, col.names = FALSE, quote = FALSE)

# Results of the function
res1 <- from_5_to_3(
  dfs = data,
  dfs_name = "tab",
  totcode = c(SEX="Total",AGE="Total", GEO="Total", ACT="Total", ECO = "PIB"),
  hrcfiles = c(ACT = hrc_act, GEO = hrc_geo, SEX = hrc_sex),
  sep_dir = TRUE,
  hrc_dir = "output",
  v1 = "ACT",
  v2 = "AGE",
  v3 = "SEX",
  v4 = "ECO"
)

res2 <- from_5_to_3(
  dfs = data,
  dfs_name = "tab",
  totcode = c(SEX="Total",AGE="Total", GEO="Total", ACT="Total", ECO = "PIB"),
  hrcfiles = c(ACT = hrc_act, GEO = hrc_geo, SEX = hrc_sex),
  sep_dir = TRUE,
  hrc_dir = "output",
  verbose = TRUE
)

InseeFrLab/rtauargus documentation built on Feb. 25, 2025, 6:32 a.m.