argusVersion

Description

returns the version and build number of a given tau-argus executable specified in argument exe.

Usage

argusVersion(exe, verbose = FALSE)

Arguments

Value

a list with two elements being the tau-argus version and the build-number.

Examples

## Not run: 
argusVersion(exe="C:\\Tau\\TauArgus.exe", verbose=TRUE)

## End(Not run)

Attacking primary suppressed cells

Description

Function [attack()] is used to compute lower and upper bounds for a given sdcProblem instance. For all calculations the current suppression pattern is used when calculating solutions of the attacker's problem.

Usage

attack(object, to_attack = NULL, verbose = FALSE, ...)

Arguments

Value

a 'data.frame' with the following columns: - 'prim_supps': index of primary suppressed cells - 'status': the original sdc-status code - 'val' the original value of the cell - ‘low': computed lower bound of the attacker’s problem - ‘up': computed upper bound of the attacker’s problem - 'protected' shows if a given cell is accordingly protected

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Examples

## Not run: 
dims <- list(
  v1 = sdcHierarchies::hier_create("tot", letters[1:4]),
  v2 = sdcHierarchies::hier_create("tot", letters[5:8])
)

N <- 150
df <- data.frame(
  v1 = sample(letters[1:4], N, replace = TRUE),
  v2 = sample(letters[5:8], N, replace = TRUE)
)

sdc <- makeProblem(data = df, dimList = dims)

# set primary suppressions
specs <- data.frame(
  v1 = c("a", "b", "a"),
  v2 = c("e", "e", "f")
)
sdc <- change_cellstatus(sdc, specs = specs, rule = "u")

# attack all primary sensitive cells
# the cells can be recomputed exactly
attack(sdc, to_attack = NULL)

# protect the table and attack again
sdc <- protectTable(sdc, method = "SIMPLEHEURISTIC")
attack(sdc, to_attack = NULL)

# attack only selected cells
attack(sdc, to_attack = c(7, 12))

## End(Not run)

perform calculations on cutList-objects depending on argument type

Description

perform calculations on cutList-objects depending on argument type

Usage

calc.cutList(object, type, input)

## S4 method for signature 'cutList,character,list'
calc.cutList(object, type, input)

Arguments

• strengthen: strenghten constraints in argument object

• checkViolation: check if a given solution violates any in argument object

• bindTogether: combine two cutList-objects

• type==strengthen: input is not used (empty list)

• type==checkViolation: input is a list of length 2

  • first element: numeric vector specifying a solution to a linear problem

  • second element: numeric vector specifying weights

• type==bindTogether: input is a list of length 1

  • first element: object of class cutList

Value

manipulated data based on argument type

• an object of class cutList if argument type matches 'strengthen' or 'bindTogether'

• a logical vector of length 1 if argument type matches 'checkViolation' with TRUE if at least one constraint is violated by the given solution

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

modify dimVar-objects depending on argument type

Description

modify dimVar-objects depending on argument type

Usage

calc.dimVar(object, type, input)

## S4 method for signature 'dimVar,character,character'
calc.dimVar(object, type, input)

Arguments

• hasDefaultCodes: calculates if a vector of codes (specified by argument input) corresponds to default codes in object

• matchCodeOrig: obtain default|standard codes for a vector of original codes specified by argument input

• matchCodeDefault: obtain original codes for a vector of default|standard codes specified by argument input

• standardize: perform standardization of level-codes (temporarily removing duplicates,..)

• requiredMinimalCodes: calculate a set of minimal codes that are required to calculate a specific (sub)total specified by argument input

Value

information from object depending on type

• a character vector if type matches 'matchCodeOrig', 'matchCodeDefault', 'standardize' or 'requiredMinimalCodes'

• a logical vector of length 1 if type matches 'hasDefaultCodes' being TRUE if argument input are default codes and FALSE otherwise

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

perform calculations on linProb-objects depending on argument type

Description

perform calculations on linProb-objects depending on argument type

Usage

calc.linProb(object, type, input)

## S4 method for signature 'linProb,character,list'
calc.linProb(object, type, input)

Arguments

• solveProblem: solve the linear problem (minimize objective function)

• fixVariables: try to fix objective variables to 0|1 based on dual costs depending on input

• type==solveProblem: a list of length 1

  • first element: character vector of length 1 specifying the solver to use.

• type==fixVariables: a list of length 3

  • first element: numeric vector specifying lower bounds for the objective variables

  • second element: numeric vector specifying upper bounds for the objective variables

  • third element: numeric vector specifying indices of primary suppressed cells

Value

manipulated data based on argument type

• list containing the solution and additional information if argument type matches 'solveProblem

• a numeric vector of indices if argument type matches 'fixVariables'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

perform calculations on multiple objects depending on argument type

Description

perform calculations on multiple objects depending on argument type

Usage

calc.multiple(type, input)

## S4 method for signature 'character,list'
calc.multiple(type, input)

Arguments

Value

manipulated data based on argument type

• list with elements 'groups', 'indices', 'strIDs', 'nrGroups' and 'nrTables' if argument type matches 'makePartitions'

• object of class linProb if argument type matches 'makeAttackerProblem'

• object of class sdcProblem if argument type matches 'calcFullProblem'

Note

internal functions/methods

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

perform calculations on problemInstance-objects depending on argument type

Description

perform calculations on problemInstance-objects depending on argument type

Usage

calc.problemInstance(object, type, input)

## S4 method for signature 'problemInstance,character,list'
calc.problemInstance(object, type, input)

Arguments

• makeMasterProblem: create the master problem that is the core of the secondary cell suppression problem

• isProtectedSolution: check if a solution violates any required (upper|lower|sliding) protection levels

• type==makeMasterProblem: input is not used (empty list)

• type==isProtectedSolution: input is a list of length 2 with elements 'input1' and 'input2'

  • element 'input1': numeric vector of calculated known lower cell bounds (from attacker's problem)

  • element 'input2': numeric vector of known upper cell bounds (from attacker's problem)

Value

information from objects of class problemInstance depending on argument type

• an object of class linProb if argument type matches 'makeMasterProblem'

• logical vector of length 1 if argument type matches 'isProtectedSolution' with TRUE if all primary suppressed cells are adequately protected, FALSE otherwise

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

perform calculations on sdcProblem-objects depending on argument type

Description

perform calculations on sdcProblem-objects depending on argument type

Usage

calc.sdcProblem(object, type, input)

## S4 method for signature 'sdcProblem,character,list'
calc.sdcProblem(object, type, input)

Arguments

• rule.freq: modify suppression status within object according to frequency suppression rule

• heuristicSolution: obtain a heuristic (greedy) solution to the problem defined by object

• cutAndBranch: solve a secondary cell suppression problem defined by object using cut and branch

• anonWorker: is used to solve the suppression problem depending on information provided with argument input

• ghmiter: solve a secondary cell suppression problem defined by object using hypercube algorithm

• preprocess: perform a preprocess procedure by trying to identify primary suppressed cells that are already protected due to other primary suppressed cells

• cellID: find index of cell defined by information provided with argument input

• finalize: create an object of class safeObj

• ghmiter.diagObj: calculate codes required to identify diagonal cells given a valid cell code - used for ghmiter-algorithm only

• ghmiter.calcInformation: calculate information for quaders identified by diagonal indices - used for ghmiter-algorithm only

• ghmiter.suppressQuader: suppress a quader based on indices

• ghmiter.selectQuader: select a quader for suppression depending on information provided with argument input - used for ghmiter-algorithm only

• ghmiter.suppressAdditionalQuader: select and suppress an additional quader (if required) based on information provided with argument input - used for ghmiter-algorithm only

• contributingIndices: calculate indices within the current problem that contribute to a given cell

• reduceProblem: reduce the problem given by object using a vector of indices

• genStructuralCuts: calculate cuts that are absolute necessary for a valid solution of the secondary cell suppression problem

• a list (typically generated using genParaObj()) specifying parameters for primary cell suppression if argument type matches 'rule.freq'

• a list if argument type matches 'heuristicSolution' having the following elements:

  • element 'aProb': an object of class linProb defining the attacker's problem

  • element 'validCuts': an object of class cutList representing a list of constraints

  • element 'solver': a character vector of length 1 specifying a solver to use

  • element 'verbose': a logical vector of length 1 setting if verbose output is desired

• a list (typically generated using genParaObj()) specifying parameters for the secondary cell suppression problem if argument type matches 'cutAndBranch', 'anonWorker', 'ghmiter', 'preprocess'

• a list of length 3 if argument type matches 'cellID' having following elements

  • first element: character vector specifying variable names that need to exist in slot 'dimInfo' of object

  • second element: character vector specifying codes for each variable that define a specific table cell

  • third element: logical vector of length 1 with TRUE setting verbosity and FALSE to turn verbose output off

• a list of length 3 if argument type matches 'ghmiter.diagObj' having following elements

  • first element: numeric vector of length 1

  • second element: a list with as many elements as dimensional variables have been specified and each element being a character vector of dimension-variable specific codes

  • third element: logical vector of length 1 defining if diagonal indices with frequency == 0 should be allowed or not

• a list of length 4 if argument type matches 'ghmiter.calcInformation' having following elements

  • first element: a list object typically generated with method calc.sdcProblem and type=='ghmiter.diagObj'

  • second element: a list with as many elements as dimensional variables have been specified and each element being a character vector of dimension-variable specific codes

  • third element: numeric vector of length 1 specifying a desired protection level

  • fourth element: logical vector of length 1 defining if quader containing empty cells should be allowed or not

• a list of length 1 if argument type matches 'ghmiter.suppressQuader' having following element

  • first element: numeric vector of indices that should be suppressed

• a list of length 2 if argument type matches 'ghmiter.selectQuader' having following elements

  • first element: a list object typically generated with method calc.sdcProblem and type=='ghmiter.calcInformation'

  • second element: a list (typically generated using genParaObj())

• a list of length 4 if argument type matches 'ghmiter.suppressAdditionalQuader' having following elements

  • first element: a list object typically generated with method calc.sdcProblem and type=='ghmiter.diagObj'

  • second element: a list object typically generated with method calc.sdcProblem and type=='ghmiter.calcInformation'

  • third element: a list object typically generated with method calc.sdcProblem and type=='ghmiter.selectQuader'

  • fourth element: a list (typically generated using genParaObj())

• a list of length 1 if argument type matches 'contributingIndices' having following element

  • first element: character vector of length 1 being an ID for which contributing indices should be calculated

• a list of length 1 if argument type matches 'reduceProblem' having following element

  • first element: numeric vector defining indices of cells that should be kept in the reduced problem

• an empty list if argument type matches 'genStructuralCuts'

Value

information from objects of class sdcProblem depending on argument type

• an object of class sdcProblem if argument type matches 'rule.freq', 'cutAndBranch', 'anonWorker', 'ghmiter', 'ghmiter.supressQuader', 'ghmiter.suppressAdditionalQuader' or 'reduceProblem'

• a numeric vector with elements being 0 or 1 if argument type matches 'heuristicSolution'

• a list if argument type matches 'preprocess' having following elements:

  • element 'sdcProblem': an object of class sdcProblem

  • element 'aProb': an object of class linProb

  • element 'validCuts': an object of class cutList

• a numeric vector of length 1 specifying the index of the cell of interest if argument type matches 'cellID'

• an object of class safeObj if argument type matches 'finalize'

• a list if argument type matches 'ghmiter.diagObj' having following elements:

  • element 'cellToProtect': character vector of length 1 defining the ID of the cell to protect

  • element 'indToProtect': numeric vector of length 1 defining the index of the cell to protect

  • element 'diagIndices': numeric vector defining indices of possible cells defining cubes

• a list containing information about each quader that could possibly be suppressed if argument type matches 'ghmiter.calcInformation'

• a list containing information about a single quader that should be suppressed if argument type matches 'ghmiter.selectQuader'

• a numeric vector with indices that contribute to the desired table cell if argument type matches 'contributingIndices'

• an object of class cutList if argument type matches 'genStructuralCuts'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

modify simpleTriplet-objects depending on argument type

Description

modify simpleTriplet-objects depending on argument type

Usage

calc.simpleTriplet(object, type, input)

## S4 method for signature 'simpleTriplet,character,list'
calc.simpleTriplet(object, type, input)

Arguments

• removeRow: remove a row with given index from object

• removeCol: remove a column with given index from object

• addRow: add a row to object

• addCol: add a column to object

• modifyRow: change specified row of object

• modifyCol: change specified column of object

• modifyCell: change specified cell of object

• bind: bind two objects of class simpleTriplet together

• type==removeRow: input is a list of length 1

  • first element: numeric vector of length 1 defining the index of the row that should be removed

• type==removeCol: input is a list of length 1

  • first element: numeric vector of length 1 defining the index of the column that should be removed

• type==addRow: input is a list of length 2

  • first element: numeric vector of column-indices

  • second element: numeric vector defining the cell-values of the row that will be added

• type==addCol: input is a list of length 2

  • first element: numeric vector of row-indices

  • second element: numeric vector defining the cell-values of the column that will be added

• type==modifyRow: input is a list of length 3

  • first element: numeric vector of length 1 specifying the the row-index of the row that will be modified

  • second element: numeric vector specifying the column-indices that should be modified

  • third element: numeric vector defining values that should be set in the given row

• type==modifyCol: input is a list of length 3

  • first element: numeric vector specifying the row-indices that should be modified

  • second element: numeric vector of length 1 specifying the the column-index of the column that will be modified

  • third element: numeric vector defining values that should be set in the given column

• type==modifyCell: input is a list of length 3

  • first element: numeric vector of length 1 defining the column-index

  • second element: numeric vector of length 1 defining the row-index

  • third element: numeric vector of length 1 holding the value that should be set in the given cell

• type==bind: input is a list of length 2

  • first element: an object of class simpleTriplet

  • second argument: is a logical vector of length 1 being TRUE if a 'rbind' or 'FALSE' if a 'cbind' should be done

Value

an object of class simpleTriplet

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Get information about specific cells

Description

Function cellInfo() can be used to query information of a single cell from a sdcProblem object. If the instance has already been protected using protectTable(), the information is retrieved from the final protected dataset, otherwise from the current state of the instance.

Usage

cell_info(object, specs, ...)

Arguments

Value

a data.frame with a row for each of the queried cells; the object contains the following columns:

• id: numeric vector of length 1 specifying the numerical index of the cell

• a column strID if object has not yet been protected

• one column for each dimensional variable

• a column freq containing the cell-frequencies

• if available, one column for each (possible) numerical value that was tabulated

• a column sdcStatus with the current sdc code

• is_primsupp: is TRUE if the cell is a primary sensitive cell

• is_secondsupp: is TRUE if the cell is a secondary suppressed cell

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Examples

# as in makeProblem() with a single primary suppression
p <- sdc_testproblem(with_supps = TRUE)
sdcProb2df(p)

# vector input
specs_vec <- c(region = "D", gender = "male")
cell_info(p, specs = specs_vec)

# data.frame input
specs_df <- data.frame(
  region = c("A", "D", "A"),
  gender = c("male", "female", "female")
)
cell_info(p, specs = specs_df)

# protect the table
p_safe <- protectTable(p, method = "SIMPLEHEURISTIC")

# re-apply
cell_info(p_safe, specs = specs_df)

Change anonymization status of a specific cell

Description

Function change_cellstatus() allows to change|modify the anonymization state of single table cells for objects of class sdcProblem.

Usage

change_cellstatus(object, specs, rule, verbose = FALSE, ...)

Arguments

Value

a sdcProblem object

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Examples

# load example-problem
# (same as example from ?makeProblem)
p <- sdc_testproblem(with_supps = FALSE)

# goal: set cells with region = "D" and gender != "total" as primary sensitive

# using a data.frame as input
specs <- data.frame(
  region = "D",
  gender = c("male", "female", "total")
)

# marking the cells as sensitive
p <- change_cellstatus(
  object = p,
  specs = specs,
  rule = "u"
)

# check
cell_info(p, specs = specs)

# using a named vector for a single cell to revert
# setting D/total as primary-sensitive

specs <- c(gender = "total", region = "D")

p <- change_cellstatus(
  object = p,
  specs = specs,
  rule = "s"
)

# and check again
cell_info(p, specs = specs)

Compute contributing units to table cells

Description

This function computes (with respect to the raw input data) the indices of all contributing units to given cells identified by ids.

Usage

contributing_indices(prob, ids = NULL)

Arguments

Value

a named ⁠list where names correspond to the given ⁠ids' and the values to the row numbers within the raw input data.

Examples

# loading test problem
p <- sdc_testproblem(with_supps = FALSE)
dt <- sdcProb2df(p, dimCodes = "original")

# question: which units contribute to cell region = "A" and gender = "female"?

# compute the id ("0102")
dt[region == "A" & gender == "female", strID]

# which indices contribute to the cell?
ids <- contributing_indices(prob = p, ids = "0101")

# check
dataObj <- get.sdcProblem(p, "dataObj")
rawData <- slot(dataObj, "rawData")
rawData[ids[["0101"]]]

# compute contributing ids for all cells
contributing_indices(p)

Create input files for tauArgus

Description

create required input-files and batch-file for tau-argus given an sdcProblem object

Usage

createArgusInput(
  obj,
  typ = "microdata",
  verbose = FALSE,
  path = getwd(),
  solver = "FREE",
  method,
  primSuppRules = NULL,
  responsevar = NULL,
  shadowvar = NULL,
  costvar = NULL,
  requestvar = NULL,
  holdingvar = NULL,
  ...
)

Arguments

Value

the filepath to the batch-file

Examples

## Not run: 
# loading micro data from sdcTable
utils::data("microdata1", package="sdcTable")
microdata1$num1 <- rnorm(mean = 100, sd = 25, nrow(microdata1))
microdata1$num2 <- round(rnorm(mean = 500, sd=125, nrow(microdata1)),2)
microdata1$weight <- sample(10:100, nrow(microdata1), replace = TRUE)

dim_region <- hier_create(root = "Total", nodes = LETTERS[1:4])

dim_region_dupl <- hier_create(root = "Total", nodes = LETTERS[1:4])
dim_region_dupl <- hier_add(dim_region_dupl, root = "B", nodes = c("b1"))
dim_region_dupl <- hier_add(dim_region_dupl, root = "D", nodes = c("d1"))

dim_gender <- hier_create(root = "Total", nodes = c("male", "female"))

dimList <- list(region = dim_region, gender = dim_gender)
dimList_dupl  <- list(region = dim_region_dupl, gender = dim_gender)
dimVarInd <- 1:2
numVarInd <- 3:5
sampWeightInd <- 6

# creating an object of class \code{\link{sdcProblem-class}}
obj <- makeProblem(
  data = microdata1,
  dimList = dimList,
  dimVarInd = dimVarInd,
  numVarInd = numVarInd,
  sampWeightInd = sampWeightInd)

# creating an object of class \code{\link{sdcProblem-class}} containing "duplicated" codes
obj_dupl <- makeProblem(
  data = microdata1,
  dimList = dimList_dupl,
  dimVarInd = dimVarInd,
  numVarInd = numVarInd,
  sampWeightInd = sampWeightInd)

## create primary suppression rules
primSuppRules <- list()
primSuppRules[[1]] <- list(type = "freq", n = 5, rg = 20)
primSuppRules[[2]] <- list(type = "p", n = 5, p = 20)
# other supported formats are:
# list(type = "nk", n=5, k=20)
# list(type = "zero", rg = 5)
# list(type = "mis", val = 1)
# list(type = "wgt", val = 1)
# list(type = "man", val = 20)

## create batchInput object
bO_md1 <- createArgusInput(
  obj = obj,
  typ = "microdata",
  path = tempdir(),
  solver = "FREE",
  method = "OPT",
  primSuppRules = primSuppRules,
  responsevar = "num1")

bO_td1 <- createArgusInput(
  obj = obj,
  typ = "tabular",
  path = tempdir(),
  solver = "FREE",
  method = "OPT")

bO_td2 <- createArgusInput(
  obj = obj_dupl,
  typ = "tabular",
  path = tempdir(),
  solver = "FREE",
  method = "OPT")

## in case CPLEX should be used, it is required to specify argument licensefile
bO_md2 <- createArgusInput(
  obj = obj,
  typ = "microdata",
  path = tempdir(),
  solver = "CPLEX",
  method = "OPT",
  primSuppRules = primSuppRules,
  responsevar = "num1",
  licensefile = "/path/to/my/cplexlicense")

## End(Not run)

Create input for jj_format

Description

This function transforms a sdcProblem object into a list that can be used as input for writeJJFormat() to write a problem in "JJ-format" to disk.

Usage

createJJFormat(x)

Arguments

Value

an input suitable for writeJJFormat()

Author(s)

Bernhard Meindl (bernhard.meindl@statistik.gv.at) and Sapphire Yu Han (y.han@cbs.nl)

Examples

# setup example problem
# microdata
utils::data("microdata1", package = "sdcTable")

# create hierarchies
dims <- list(
  region = sdcHierarchies::hier_create(root = "Total", nodes = LETTERS[1:4]),
  gender = sdcHierarchies::hier_create(root = "Total", nodes = c("male", "female")))

# create a problem instance
p <- makeProblem(
  data = microdata1,
  dimList = dims,
  numVarInd = "val")

# create suitable input for `writeJJFormat`
inp <- createJJFormat(p); inp

# write files to disk
# frequency table by default
writeJJFormat(
  x = inp,
  path = file.path(tempdir(), "prob_freqs.jj"),
  overwrite = TRUE
)

# or using the numeric variable `val` previously specified
writeJJFormat(
  x = inp,
  tabvar = "val",
  path = file.path(tempdir(), "prob_val.jj"),
  overwrite = TRUE
)

Create input for RegSDC/other Tools

Description

This function transforms a sdcProblem object into an object that can be used as input for RegSDC::SuppressDec (among others).

Usage

createRegSDCInput(x, chk = FALSE)

Arguments

Value

an list with the following elements:

• mat: linear combinations depending on inner-cells of the given problem instance.

• y: a 1-column matrix containing the frequencies of inner cells

• z: a 1-column matrix containing the frequencies of all cells

• z_supp: a 1-column matrix containing the frequencies of all cells but suppressed cells have a value of NA

• info: a data.frame with the following columns:

  • cell_id: internal cell-id used in sdcTable

  • is_innercell: a binary indicator if the cell is an internal cell (TRUE) or a (sub)total (FALSE)

Author(s)

Bernhard Meindl (bernhard.meindl@gmail.com)

Examples

## Not run: 
utils::data("microdata1", package = "sdcTable")
head(microdata1)

# define the problem
dim_region <- hier_create(root = "total", nodes = sort(unique(microdata1$region)))
dim_gender <- hier_create(root = "total", nodes = sort(unique(microdata1$gender)))

prob <- makeProblem(
  data = microdata1,
  dimList = list(region = dim_region, gender = dim_gender),
  freqVarInd = NULL
)

# suppress some cells
prob <- primarySuppression(prob, type = "freq", maxN = 15)

# compute input for RegSDC-package
inp_regsdc <- createRegSDCInput(x = prob, chk = TRUE)

# estimate innner cells based on linear dependencies
res_regsdc <- RegSDC::SuppressDec(
  x = as.matrix(inp_regsdc$x),
  z = inp_regsdc$z_supp,
  y = inp_regsdc$y)[, 1]

# check if inner cells are all protected
df <- data.frame(
  freqs_orig = inp_regsdc$z[inp_regsdc$info$is_innercell == TRUE, ],
  freqs_supp = inp_regsdc$z_supp[inp_regsdc$info$is_innercell == TRUE, ],
  regsdc = res_regsdc
)

subset(df, df$regsdc == df$freqs_orig & is.na(freqs_supp))


## End(Not run)

Create a hierarchy

Description

create_node() is defunct, please use sdcHierarchies::hier_create()

add_nodes() is defunct, please use sdcHierarchies::hier_add()

delete_nodes() is defunct, please use sdcHierarchies::hier_delete()

rename_node() is defunct, please use sdcHierarchies::hier_rename()

cellInfo() is defunct, please use [cell_info()]

changeCellStatus() is defunct, please use [change_cellstatus()]

Usage

create_node(...)

add_nodes(...)

delete_nodes(...)

rename_node(...)

cellInfo(...)

changeCellStatus(...)

S4 class describing a cutList-object

Description

An object of class cutList holds constraints that can be extracted and used as for objects of class linProb-class. An object of class cutList consists of a constraint matrix (slot con), a vector of directions (slot direction) and a vector specifying the right hand sides of the constraints (slot rhs).

Details

slot con:

an object of class simpleTriplet-class specifying the constraint matrix of the problem

slot direction:

a character vector holding the directions of the constraints, allowed values are:

• ==: equal

• <: less

• >: greater

• <=: less or equal

• >=: greater or equal

slot rhs:

numeric vector holding right hand side values of the constraints

Note

objects of class cutList are dynamically generated (and removed) during the cut and branch algorithm when solving the secondary cell suppression problem

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

S4 class describing a dataObj-object

Description

This class models a data object containing the 'raw' data for a given problem as well as information on the position of the dimensional variables, the count variable, additional numerical variables, weights or sampling weights within the raw data. Also slot 'isMicroData' shows if slow 'rawData' consists of microdata (multiple observations for each cell are possible, isMicroData==TRUE) or if data have already been aggregated (isMicroData==FALSE)

Details

slot rawData:

list with each element being a vector of either codes of dimensional variables, counts, weights that should be used for secondary cell suppression problem, numerical variables or sampling weights.

slot dimVarInd:

numeric vector (or NULL) defining the indices of the dimensional variables within slot 'rawData'

slot freqVarInd:

numeric vector (or NULL) defining the indices of the frequency variables within slot 'rawData'

slot numVarInd:

numeric vector (or NULL) defining the indices of the numerical variables within slot 'rawData'

slot weightVarInd:

numeric vector (or NULL) defining the indices of the variables holding weights within slot 'rawData'

slot sampWeightInd:

numeric vector (or NULL) defining the indices of the variables holding sampling weights within slot 'rawData'

slot isMicroData:

logical vector of length 1 (or NULL) that is TRUE if slot 'rawData' are microData and FALSE otherwise

Note

objects of class dataObj are input for slot dataObj in class sdcProblem

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

S4 class describing a dimInfo-object

Description

An object of class dimInfo holds all necessary information about the dimensional variables defining a hierarchical table that needs to be protected.

Details

slot dimInfo:

a list (or NULL) with all list elements being objects of class dimVar

slot strID:

a character vector (or NULL) defining IDs that identify each table cell. The ID's are based on (default) codes of the dimensional variables defining a cell.

slot strInfo:

a list object (or NULL) with each list element being a numeric vector of length 2 defining the start and end-digit that is allocated by the i-th dimensional variable in ID-codes available in slot strID

slot vNames:

a character vector (or NULL) defining the variable names of the dimensional variables defining the table structure

slot posIndex:

a numeric vector (or NULL) holding the position of the dimensional variables within slot rawData of class dataObj

Note

objects of class dimInfo are input for slots in classes sdcProblem and safeObj

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

S4 class describing a dimVar-object

Description

An object of class dimVar holds all necessary information about a single dimensional variable such as original and standardized codes, the level-structure, the hierarchical structure, codes that may be (temporarily) removed from building the complete hierarchy (dups) and their corresponding codes that correspond to these duplicated codes.

Details

slot codesOriginal:

a character vector (or NULL) holding original variable codes

slot codesDefault:

a character vector (or NULL) holding standardized codes

slot codesMinimal:

a logical vector (or NULL) defining if a code is required to build the complete hierarchy or not (then the code is a (sub)total)

slot vName:

character vector of length 1 (or NULL) defining the variable name of the dimensional variable

slot levels:

a numeric vector (or NULL) defining the level structure. For each code the corresponding level is listed with the grand-total always having level==1

slot structure:

a numeric vector (or NULL) with length of the total number of levels. Each element shows how many digits the i-th level allocates within the standardized codes (note: level 1 always allocates exactly 1 digit in the standardized codes)

slot dims:

a list (or NULL) defining the hierarchical structure of the dimensional variable. Each list-element is a character vector with elements available in slot codesDefault and the first element always being a (sub)total and the remaining elements being the codes that contribute to the (sub)total

slot dups:

character vector (or NULL) having showing original codes that are duplicates in the hierarchy and can temporarily removed when building a table with this dimensional variable

slot dupsUp:

character vector (or NULL) with original codes that are the corresponding upper-levels to the codes that may be removed because they are duplicates and that are listed in slot dups

Note

objects of class dimVar form the base for elements in slot dimInfo of class dimInfo.

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

query cutList-objects depending on argument type

Description

query cutList-objects depending on argument type

Usage

get.cutList(object, type)

## S4 method for signature 'cutList,character'
get.cutList(object, type)

Arguments

• constraints: constraint matrix of object

• direction: directions of the constraints

• rhs: right hand side of the constraints

• nrConstraints: total number of constraints

Value

information from objects of class cutList depending on argument type

• object of class simpleTriplet if argument type matches 'constraints'

• character vector if argument type matches 'direction'

• numeric vector if argument type matches 'objective' or 'rhs'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

query dataObj-objects depending on argument type

Description

query dataObj-objects depending on argument type

Usage

get.dataObj(object, type)

## S4 method for signature 'dataObj,character'
get.dataObj(object, type)

Arguments

• rawData: raw input data

• dimVarInd: indices of dimensional variables

• freqVarInd: index of frequency variable

• numVarInd: indices of numerical variables

• weightVarInd index of weight variable

• sampWeightInd index of variable holding sampling weights

• isMicroData does object consist of microdata?

• numVarNames variable names of numerical variables

• freqVarName variable name of frequency variable

• varName variable names of dimensional variables

Value

information from objects of class dataObj depending on argument type

• a list if argument type matches 'rawData'

• numeric vector if argument type matches 'dimVarInd', 'freqVarInd', 'numVarInd', 'weightVarInd' or 'sampWeightInd'

• character vector if argument type matches 'numVarNames', 'freqVarName' or 'varName'

• logical vector of length 1 if argument type matches 'isMicroData'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

query dimInfo-objects depending on argument type

Description

query dimInfo-objects depending on argument type

Usage

get.dimInfo(object, type)

## S4 method for signature 'dimInfo,character'
get.dimInfo(object, type)

Arguments

• strInfo: info on how many digits in the default codes ach dimensional variable allocates

• dimInfo: a list object with each slot containing an object of class dimVar

• varName: variable names

• strID: character vector of ID's defining table cells

• posIndex vector showing the index of the elements of dimInfo in the underlying data

Value

information from objects of class dimInfo depending on argument type

• a list (or NULL) if argument type matches 'strInfo', 'dimInfo'

• numeric vector (or NULL) if argument type matches 'posIndex'

• character vector (or NULL) if argument type matches 'varName' or 'strID'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

query dimVar-objects depending on argument type

Description

query dimVar-objects depending on argument type

Usage

get.dimVar(object, type)

## S4 method for signature 'dimVar,character'
get.dimVar(object, type)

Arguments

• varName: variable name of the variable from which object was calculated

• codesOriginal: original codes (as specified by the user)

• codesDefault: calculated, default codes

• codesMinimal: all codes required to calculate the complete hierarchy (no sub-totals)

• levels: level-structure of the dimensional variable

• structure: vector showing how many digits in the default codes are required for each level

• dims: list showing the complete hierarchy of the dimensional variable

• dups: vector of duplicated codes

• dupsUp: vector of codes that are the 'upper' levels to which the codes in dups correspond

• hasDuplicates: does the dimensional variable has codes that can be (temporarily) removed

• nrLevels: the total number of levels of a dimensional variable

• minimalCodesDefault: the standardized codes of the minimal set of required level-codes

Value

information from objects of class dataObj depending on argument type

• a list if argument type matches 'dims'

• numeric vector if argument type matches 'levels' or 'nrLevels'

• character vector if argument type matches 'codesOriginal', 'codesDefault', 'vName', 'dups', 'dupsUp' or 'minimalCodesDefault'

• logical vector of length 1 if argument type matches 'hasDuplicates'

• a logical vector if argument type matches 'codesMinimal'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

query linProb-objects depending on argument type

Description

query linProb-objects depending on argument type

Usage

get.linProb(object, type)

## S4 method for signature 'linProb,character'
get.linProb(object, type)

Arguments

• constraints: constraint matrix of object linProb

• direction: directions of the constraints

• rhs: right hand side of the constraints

• objective: objective function

• types: types of the objective variables

• bounds: bounds of the objective variables

Value

information from objects of class linProb depending on type

• an object of class simpleTriplet if type matches 'constraints'

• a character vector if type matches 'direction' or 'types'

• a numeric vector if type matches 'objective' or 'rhs'

• a list with elements 'lower' and 'upper' if type matches 'bounds'

  • element 'lower': a list with the first element containing indices and the second element containing corrsponding lower bounds

  • element 'upper': a list with the first element containing indices and the second element containing corrsponding upper bounds

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

query problemInstance-objects depending on argument type

Description

query problemInstance-objects depending on argument type

Usage

get.problemInstance(object, type)

## S4 method for signature 'problemInstance,character'
get.problemInstance(object, type)

Arguments

• strID: vector of unique IDs for each table cell

• nrVars: total number of table cells

• freq: vector of frequencies

• w: a vector of weights used in the linear problem (or NULL)

• numVars: a list containing numeric vectors containing values for numerical variables for each table cell (or NULL)

• sdcStatus: a vector containing the suppression state for each cell (possible values are 'u': primary suppression, 'x': secondary suppression, 'z': forced for publication, 's': publishable cell, 'w': dummy cells that are considered only when applying the simple greedy heuristic to protect the table)

• lb: lower bound assumed to be known by attackers for each table cell

• ub: upper bound assumed to be known by attackers for each table cell

• LPL: lower protection level required to protect table cells

• UPL: upper protection level required to protect table cells

• SPL: sliding protection level required to protect table cells

• primSupps: vector of indices of primary sensitive cells

• secondSupps: vector of indices of secondary suppressed cells

• forcedCells: vector of indices of cells that must not be suppressed

• hasPrimSupps: shows if object has primary suppressions or not

• hasSecondSupps: shows if object has secondary suppressions or not

• hasForcedCells: shows if object has cells that must not be suppressed

• weight: gives weight that is used the suppression procedures

• suppPattern: gives the current suppression pattern

Value

information from objects of class dataObj depending on argument type

• a list (or NULL) if argument type matches 'numVars'

• numeric vector if argument type matches 'freq', 'lb', 'ub', 'LPL', 'UPL', 'SPL', 'weight', 'suppPattern'

• numeric vector (or NULL) if argument type matches 'w', 'primSupps', 'secondSupps', 'forcedCells'

• character vector if argument type matches 'strID', 'sdcStatus', ”

• logical vector of length 1 if argument type matches 'hasPrimSupps', 'hasSecondSupps', 'hasForcedCells'

• numerical vector of length 1 if argument type matches 'nrVars'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

query sdcProblem-objects depending on argument type

Description

query sdcProblem-objects depending on argument type

Usage

get.sdcProblem(object, type)

## S4 method for signature 'sdcProblem,character'
get.sdcProblem(object, type)

Arguments

• dataObj: a list containing the (raw) input data

• problemInstance: return the current problem instance

• partition: a list containing information on the subtables that are required to be protected as well as information on the processing order of the subtables

• dimInfo: information on the variables defining the hierarchical table

• indicesDealtWith: a set of indices that have already been dealt with during the protection algorithmus

• startI: current level at which subtables need to be protected (useful when restarting HITAS|HYPERCUBE)

• startJ: current number of the subtable within a given level that needs to be protected (useful when restarting HITAS|HYPERCUBE)

• innerAndMarginalCellInfo: for a given problem, get indices of inner- and marginal table cells

Value

information from objects of class sdcProblem depending on argument type

• an object of class dataObj (or NULL) if type matches 'dataObj'

• an object of class problemInstance (or NULL) if type matches 'problemInstance'

• a list (or NULL) if argument type matches 'partition' containing the following elements:

  • element 'groups': list with each list-element being a character vector specifying a specific level-group

  • element 'indices': list with each list-element being a numeric vector defining indices of a subtable

  • element 'strIDs': list with each list-element being a character vector defining IDs of a subtable

  • element 'nrGroups': numeric vector of length 1 defining the total number of groups that have to be considered

  • element 'nrTables': numeric vector of length 1 defining the total number of subtables that have to be considered

• a list (or NULL) if argument type matches 'innerAndMarginalCellInfo' containing the following elements:

  • element 'innerCells': character vector specifying ID's of inner cells

  • element 'totCells': character vector specifying ID's of marginal cells

  • element 'indexInnerCells': numeric vector specifying indices of inner cells

  • element 'indexTotCells': numeric vector specifying indices of marginal cells

• an object of class dimInfo (or NULL) if type matches 'dimInfo'

• numeric vector of length 1 if argument type matches 'startI' or 'startJ'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

query simpleTriplet-objects depending on argument type

Description

query simpleTriplet-objects depending on argument type

Usage

get.simpleTriplet(object, type, input)

## S4 method for signature 'simpleTriplet,character,list'
get.simpleTriplet(object, type, input)

Arguments

• rowInd: extract all row-indices

• colInd: extract all column-indices

• values: extract all values

• nrRows: return the number of rows of the input object

• nrCols: return the number of columns of the input object

• nrCells: return the number of cells (different from 0!)

• duplicatedRows: return a numeric vector showing indices of duplicated rows

• transpose: transpose input object and return the transposed matrix

• getRow: return a specific row of input object

• getCol: return a specific column of input object

• type == 'getRow': input is a list of length 1

  • first element: numeric vector of length 1 defining index of row that is to be returned

• type == 'getCol': input is a list of length 1

  • first element: numeric vector of length 1 defining index of column that is to be returned

• else: input is not used at all (empty list)

Value

information from object depending on type

• a numeric vector if type matches 'rowInd', 'colInd', 'values', 'nrRows', 'nrCols', 'nrCells' or 'duplicatedRows'

• an object of class simpleTriplet if type matches 'transpose', 'getRow' or 'getCol'

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Retrieve information in sdcProblem or problemInstance objects

Description

Function getInfo() is used to extract values from sdcProblem or problemInstance objects

Usage

getInfo(object, type)

Arguments

Value

manipulated data depending on arguments object and type

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Examples

# define an example problem with two hierarchies
p <- sdc_testproblem(with_supps = FALSE)

# apply primary suppression
p <- primarySuppression(p, type = "freq", maxN = 3)

# `p` is an `sdcProblem` object
print(class(p))

for (slot in c("lb", "ub", "LPL", "SPL", "UPL", "sdcStatus",
  "freq", "strID", "numVars", "w")) {
  message("slot: ", shQuote(slot))
  print(getInfo(p, type = slot))
}

# protect the cell and extract results
p_protected <- protectTable(p, method = "SIMPLEHEURISTIC")
for (slot in c("finalData", "nrNonDuplicatedCells", "nrPrimSupps",
  "nrSecondSupps", "nrPublishableCells", "suppMethod")) {
  message("slot: ", shQuote(slot))
  print(getInfo(p_protected, type = slot))
}

Query information from protected problem instances

Description

get_safeobj() allows to extract information from protected sdcProblem instances.

Usage

get_safeobj(object, type, ...)

Arguments

Value

the required information.

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

initialize cutList-objects depending on argument type

Description

initialize cutList-objects depending on argument type

Usage

init.cutList(type, input)

## S4 method for signature 'character,list'
init.cutList(type, input)

Arguments

• empty: create an empty cutList-object

• singleCut: create a cutList-object with exactly one constraint

• multipleCuts: create a cutList-object with more than one constraint

• type==empty: input is not used (empty list)

• type==singleCut: input is a list of length 3

  • first element: numeric vector specifying a values for the row of the constraint matrix that must be created

  • second element: character vector of length 1 specifying the direction

  • third element: numeric vector of length 1 specifying the right hand side of the constraint

• type==multipleCuts: input is a list of length 3

  • first element: object of class matrix

  • second element: character vector specifying the direction of the constraints

  • third element: numeric vector specifying the right hand side of the constraints

Value

an object of class cutList

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

initialize dataObj-objects

Description

initialize dataObj-objects

Usage

init.dataObj(input)

## S4 method for signature 'list'
init.dataObj(input)

Arguments

• element 'inputData': a list object holding data

• element 'dimVarInd': index (within inputData) of variables that define the table to protect

• element 'freqVarInd': index (within inputData) of variable holding frequencies

• element 'numVarInd' index (within inputData) of numerical variables (or NULL)

• element 'weightInd': index (within inputData) of variable holding weights (or NULL)

• element 'sampWeightInd': index (within inputData) of variable holding sampling weights (or NULL)

• element 'isMicroData': logical vector of length 1

Value

an object of class dataObj

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

initialize dimVar-object

Description

initialize dimVar-object

Usage

init.dimVar(input)

## S4 method for signature 'list'
init.dimVar(input)

Arguments

• first element: either an object of class 'matrix' or a data.frame or a link to a file. The input data need to be in a specific format (2 columns) with the first column defining the level-structure and the second column defining the level-codes.

• second element: a character vector of length 1 specifying a variable name

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

initialize simpleTriplet-objects depending on argument type

Description

init.simpleTriplet should be used to create objects of class simpleTriplet. It is possible to create an object from class simpleTriplet from an existing matrix (using type=='simpleTriplet'). A positive (or negative) identity matrix stored as an object of class simpleTriplet can be created by specifying type=='simpleTripletDiag'.

Usage

init.simpleTriplet(type, input)

## S4 method for signature 'character,list'
init.simpleTriplet(type, input)

Arguments

• simpleTriplet: a simple triplet matrix

• simpleTripletDiag: identity matrix

• type == 'simpleTriplet': input is a list of length 1

  • first element: object of class 'matrix'

• type == 'simpleTripletDiag': input is a list of length 2

  • first element: numeric vector of length 1 defining the desired number of rows of the identiy matrix

  • second element: logical vector of length 1 being TRUE if a positive and FALSE if a negative identity matrix should be returned

Value

an object of class simpleTriplet

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

S4 class describing a linProb-object

Description

An object of class linProb defines a linear problem given by the objective coefficients (slot objective), a constraint matrix (slot constraints), the direction (slot direction) and the right hand side (slot rhs) of the constraints. Also, allowed lower (slot boundsLower) and upper (slot boundsUpper) bounds of the variables as well as its types (slot types) are specified.

Details

slot objective:

a numeric vector holding coefficients of the objective function

slot constraints:

an object of class simpleTriplet-class specifying the constraint matrix of the problem

slot direction:

a character vector holding the directions of the constraints, allowed values are:

• ==: equal

• <: less

• >: greater

• <=: less or equal

• >=: greater or equal

slot rhs:

numeric vector holding right hand side values of the constraints

slot boundsLower:

a numeric vector holding lower bounds of the objective variables

slot boundsUpper:

a numeric vector holding upper bounds of the objective variables

slot types:

a character vector specifying types of the objective variables, allowed types are:

• C: binary

• B: continuous

• I: integer

Note

when solving the problems in the procedure, minimization of the objective is performed.

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Create a problem instance

Description

Function makeProblem() is used to create sdcProblem objects.

Usage

makeProblem(
  data,
  dimList,
  dimVarInd = NULL,
  freqVarInd = NULL,
  numVarInd = NULL,
  weightInd = NULL,
  sampWeightInd = NULL
)

Arguments

Value

a sdcProblem object

Author(s)

Bernhard Meindl

Examples

# loading micro data
utils::data("microdata1", package = "sdcTable")

# we can observe that we have a micro data set consisting
# of two spanning variables ('region' and 'gender') and one
# numeric variable ('val')

# specify structure of hierarchical variable 'region'
# levels 'A' to 'D' sum up to a Total
dim.region <- data.frame(
 levels=c('@','@@','@@','@@','@@'),
 codes=c('Total', 'A','B','C','D'),
 stringsAsFactors=FALSE)

# specify structure of hierarchical variable 'gender'
# using create_node() and add_nodes() (see ?manage_hierarchies)
dim.gender <- hier_create(root = "Total", nodes = c("male", "female"))
hier_display(dim.gender)

# create a named list with each element being a data-frame
# containing information on one dimensional variable and
# the names referring to variables in the input data
dimList <- list(region = dim.region, gender = dim.gender)

# third column containts a numeric variable
numVarInd <- 3

# no variables holding counts, numeric values, weights or sampling
# weights are available in the input data
# creating an problem instance using numeric indices
p1 <- makeProblem(
  data = microdata1,
  dimList = dimList,
  numVarInd = 3 # third variable in `data`
)

# using variable names is also possible
p2 <- makeProblem(
  data = microdata1,
  dimList = dimList,
  numVarInd = "val"
)

# what do we have?
print(class(p1))

# have a look at the data
df1 <- sdcProb2df(p1, addDups = TRUE,
  addNumVars = TRUE, dimCodes = "original")
df2 <- sdcProb2df(p2, addDups=TRUE,
  addNumVars = TRUE, dimCodes = "original")
print(df1)

identical(df1, df2)

Synthetic Microdata (1)

Description

A 'data.frame' used for examples and problem-generation in various examples.

Usage

data(microdata1)

Format

a 'data.frame' with '100' rows and variables 'region', 'gender' and 'val'.

Examples

utils::data("microdata1", package = "sdcTable")
head(microdata1)

Synthetic Microdata (2)

Description

Example microdata used for example in [protect_linked_tables()].

Usage

data(microdata2)

Format

a 'data.frame' with '100' observations containing variables 'region', 'gender', 'ecoOld', 'ecoNew' and 'numVal'.

Examples

utils::data("microdata2", package = "sdcTable")
head(microdata2)

Apply primary suppression

Description

Function primarySuppression() is used to identify and suppress primary sensitive table cells in sdcProblem objects. Argument type allows to select a rule that should be used to identify primary sensitive cells. At the moment it is possible to identify and suppress sensitive table cells using the frequency-rule, the nk-dominance rule and the p-percent rule.

Usage

primarySuppression(object, type, ...)

Arguments

Details

since versions ⁠>= 0.29⁠ it is no longer possible to specify underlying variables for dominance rules ("p", "pq" or "nk") by index; these variables must be set by name using argument numVarName.

Value

a sdcProblem object

Note

the nk-dominance rule, the p-percent rule and the pq-rule can only be applied if micro data have been used as input data to function makeProblem()

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Examples

# load micro data
utils::data("microdata1", package = "sdcTable")

# load problem (as it was created in the example in ?makeProblem
p <- sdc_testproblem(with_supps = FALSE)

# we have a look at the frequency table by gender and region
xtabs(rep(1, nrow(microdata1)) ~ gender + region, data = microdata1)

# 2 units contribute to cell with region=='A' and gender=='female'
# --> this cell is considered sensitive according the the
# freq-rule with 'maxN' equal to 2!
p1 <- primarySuppression(
  object = p,
  type = "freq",
  maxN = 2
)

# we can also apply a p-percent rule with parameter "p" being 30 as below.
# This is only possible if we are dealing with micro data and we also
# have to specify the name of a numeric variable.
p2 <- primarySuppression(
  object = p,
  type = "p",
  p = 30,
  numVarName = "val"
)

# looking at anonymization states we see, that one cell is primary
# suppressed (sdcStatus == "u")
# the remaining cells are possible candidates for secondary cell
# suppression (sdcStatus == "s") given the frequency rule with
# parameter "maxN = 2".
#
# Applying the p-percent rule with parameter 'p = 30' resulted in
# two primary suppressions.
data.frame(
  p1_sdc = getInfo(p1, type = "sdcStatus"),
  p2_sdc = getInfo(p2, type = "sdcStatus")
)

print dimVar-class objects

Description

print dimVar-class objects in a resonable way

Usage

## S4 method for signature 'dimVar'
print(x, ...)

Arguments

print objects of class sdcProblem-class.

Description

print some useful information instead of just displaying the entire object (which may be large)

Usage

## S4 method for signature 'sdcProblem'
print(x, ...)

Arguments

S4 class describing a problemInstance-object

Description

An object of class problemInstance holds the main information that is required to solve the secondary cell suppression problem.

Details

slot strID:

a character vector (or NULL) of ID's identifying table cells

slot Freq:

a numeric vector (or NULL) of counts for each table cell

slot w:

a numeric vector (or NULL) of weights that should be used when solving the secondary cell suppression problem

slot numVars:

a list (or NULL) with each element being a numeric vector holding values of specified numerical variables for each table cell

slot lb:

numeric vector (or NULL) holding assumed lower bounds for each table cell

slot ub:

numeric vector (or NULL) holding assumed upper bounds for each table cell

slot LPL:

numeric vector (or NULL) holding required lower protection levels for each table cell

slot UPL:

numeric vector (or NULL) holding required upper protection levels for each table cell

slot SPL:

numeric vector (or NULL) holding required sliding protection levels for each table cell

slot sdcStatus:

character vector (or NULL) holding the current anonymization state for each cell.

• z: cell is forced to be published and must not be suppressed

• u: cell has been primary suppressed

• x: cell is a secondary suppression

• s: cell can be published

Note

objects of class problemInstance are used as input for slot problemInstance in class sdcProblem

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Protect two tables with common cells

Description

protect_linked_tables() can be used to protect tables that have common cells. It is of course required that after the anonymization process has finished, all common cells have the same anonymization state in both tables.

Usage

protectLinkedTables(
  objectA,
  objectB,
  commonCells,
  method = "SIMPLEHEURISTIC",
  ...
)

protect_linked_tables(x, y, common_cells, method = "SIMPLEHEURISTIC", ...)

Arguments

Value

a list elements x and y containing protected sdcProblem objects

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

See Also

protectTable()

Examples

## Not run: 
# load micro data for further processing
utils::data("microdata2", package = "sdcTable")

# table1: defined by variables 'gender' and 'ecoOld'
md1 <- microdata2[,c(2,3,5)]

# table2: defined by variables 'region', 'gender' and 'ecoNew'
md2 <- microdata2[,c(1,2,4,5)]

# we need to create information on the hierarchies
# variable 'region': exists only in md2
d_region <- hier_create(root = "Tot", nodes = c("R1", "R2"))

# variable 'gender': exists in both datasets
d_gender <- hier_create(root = "Tot", nodes = c("m", "f"))

# variable 'eco1': exists only in md1
d_eco1 <- hier_create(root = "Tot", nodes = c("A", "B"))
d_eco1 <- hier_add(d_eco1, root = "A", nodes = c("Aa", "Ab"))
d_eco1 <- hier_add(d_eco1, root = "B", nodes = c("Ba", "Bb"))

# variable 'ecoNew': exists only in md2
d_eco2 <- hier_create(root = "Tot", nodes = c("C", "D"))
d_eco2 <- hier_add(d_eco2, root = "C", nodes = c("Ca", "Cb", "Cc"))
d_eco2 <- hier_add(d_eco2, root = "D", nodes = c("Da", "Db", "Dc"))

# creating objects holding information on dimensions
dl1 <- list(gender = d_gender, ecoOld = d_eco1)
dl2 <- list(region = d_region, gender = d_gender, ecoNew = d_eco2)

# creating input objects for further processing.
# For details, see ?makeProblem.
p1 <- makeProblem(
  data = md1,
  dimList = dl1,
  dimVarInd = 1:2,
  numVarInd = 3)

p2 <- makeProblem(
  data = md2,
  dimList = dl2,
  dimVarInd = 1:3,
  numVarInd = 4)

# the cell specified by gender == "Tot" and ecoOld == "A"
# is one of the common cells! -> we mark it as primary suppression
p1 <- change_cellstatus(
  object = p1,
  specs = data.frame(gender = "Tot", ecoOld = "A"),
  rule = "u",
  verbose = FALSE)

# the cell specified by region == "Tot" and gender == "f" and ecoNew == "C"
# is one of the common cells! -> we mark it as primary suppression
p2 <- change_cellstatus(
  object = p2,
  specs = data.frame(region = "Tot", gender = "f", ecoNew = "C"),
  rule = "u",
  verbose = FALSE)

# specifying input to define common cells
common_cells <- list()

# variable "gender"
common_cells$v.gender <- list()
common_cells$v.gender[[1]] <- "gender" # variable name in "p1"
common_cells$v.gender[[2]] <- "gender" # variable name in "p2"

# "gender" has equal characteristics on both datasets -> keyword "ALL"
common_cells$v.gender[[3]] <- "ALL"

# variables: "ecoOld" and "ecoNew"
common_cells$v.eco <- list()
common_cells$v.eco[[1]] <- "ecoOld" # variable name in "p1"
common_cells$v.eco[[2]] <- "ecoNew" # variable name in "p2"

# vector of common characteristics:
# "A" and "B" in variable "ecoOld" in "p1"
common_cells$v.eco[[3]] <- c("A", "B")

# correspond to codes "C" and "D" in variable "ecoNew" in "p2"
common_cells$v.eco[[4]] <- c("C", "D")

# protect the linked data
result <- protect_linked_tables(
  x = p1,
  y = p2,
  common_cells = common_cells,
  verbose = TRUE)

# having a look at the results
result_tab1 <- result$x
result_tab2 <- result$y
summary(result_tab1)
summary(result_tab2)

## End(Not run)

Protecting sdcProblem objects

Description

Function protectTable() is used to protect primary sensitive table cells (that usually have been identified and set using primarySuppression()). The function protects primary sensitive table cells according to the method that has been chosen and the parameters that have been set. Additional parameters that are used to control the protection algorithm are set using parameter ....

Usage

protectTable(object, method, ...)

Arguments

Details

The implemented methods may have bugs that yield in not-fully protected tables. Especially the usage of "OPT", "HITAS" and "HYPERCUBE" in production is not suggested as these methods may eventually be removed completely. In case you encounter any problems, please report it or use Tau-Argus (https://research.cbs.nl/casc/tau.htm).

Value

an safeObj object

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Examples

## Not run: 
# load example-problem with with a single primary suppression
# (same as example from ?primarySuppression)
p <- sdc_testproblem(with_supps = TRUE)

# protect the table using the 'GAUSS' algorithm with verbose output
res1 <- protectTable(p, method = "GAUSS", verbose = TRUE)
res1

# protect the table using the 'HITAS' algorithm with verbose output
res2 <- protectTable(p, method = "HITAS", verbose = TRUE, useC = TRUE)
res2

# protect using the heuristic algorithm
res3 <- protectTable(p, method = "SIMPLEHEURISTIC")
res3

# protect using the old implmentation of the heuristic algorithm
# used in sdcTable versions <0.32
res4 <- protectTable(p, method = "SIMPLEHEURISTIC_OLD")
res4

# looking at the final table with result suppression pattern
print(getInfo(res1, type = "finalData"))

## End(Not run)

runArgusBatchFile

Description

allows to run batch-files for tau argus given the path to an executable of argus. The provided batch input files can either be created using function createArgusInput or can be arbitrarily created. In the latter case, argument obj should not be specified and not output is returned, the script is just executed in tau-argus.

Usage

runArgusBatchFile(
  obj = NULL,
  batchF,
  exe = "C:\\Tau\\TauArgus.exe",
  batchDataDir = NULL,
  verbose = FALSE
)

Arguments

Value

a data.table containing the protected table or an error in case the batch-file was not solved correctly if the batch-file was created using sdcTable (argument obj) was specified. In case an arbitrarily batch-file has been run, NULL is returned.

Note

in case a custom batch-file is used as input (e.g obj is NULL), this functions does currently not try to read in any tables to the system.

S4 class describing a safeObj-object

Description

Objects of class safeObj are the final result after protection a tabular structure. After a successful run of protectTable an object of this class is generated and returned. Objects of class safeObj contain a final, complete data set (slot finalData) that has a column showing the anonymization state of each cell and the complete information on the dimensional variables that have defined the table that has been protected (slot dimInfo). Also, the number of non-duplicated table cells (slot nrNonDuplicatedCells) is returned along with the number of primary (slot nrPrimSupps) and secondary (slot nrSecondSupps) suppressions. Furthermore, the number of cells that can be published (slot nrPublishableCells) and the algorithm that has been used to protect the data (slot suppMethod) is returned.

Details

slot finalData:

a data.frame (or NULL) featuring columns for each variable defining the table (with their original codes), the cell counts and values of any numerical variables and the anonymization status for each cell with

• s, z: cell can be published

• u: cell is a primary sensitive cell

• x: cell was selected as a secondary suppression

slot dimInfo:

an object of class dimInfo-class holding all information on variables defining the table

slot nrNonDuplicatedCells:

numeric vector of length 1 (or NULL) showing the number of non-duplicated table cells. This value is different from 0 if any dimensional variable features duplicated codes. These codes have been re-added to the final dataset.

slot nrPrimSupps:

numeric vector of length 1 (or NULL) showing the number of primary suppressed cells

slot nrSecondSupps:

numeric vector of length 1 (or NULL) showing the number of secondary suppressions

slot nrPublishableCells:

numeric vector of length 1 (or NULL) showing the number of cells that may be published

slot suppMethod:

character vector of length 1 holding information on the protection method

Note

objects of class safeObj are returned after the function protectTable has finished.

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Transform a problem instance

Description

sdcProb2df() returns a data.table given an sdcProblem input object.

Usage

sdcProb2df(obj, addDups = TRUE, addNumVars = FALSE, dimCodes = "both")

Arguments

Value

a data.table containing information about all cells of the given problem

Examples

# loading micro data
utils::data("microdata1", package = "sdcTable")

# we can observe that we have a micro data set consisting
# of two spanning variables ('region' and 'gender') and one
# numeric variable ('val')

# specify structure of hierarchical variable 'region'
# levels 'A' to 'D' sum up to a Total
dim.region <- data.frame(
 levels=c('@','@@','@@','@@','@@'),
 codes=c('Total', 'A','B','C','D'),
 stringsAsFactors=FALSE)

# specify structure of hierarchical variable 'gender'
# using create_node() and add_nodes() (see ?manage_hierarchies)
dim.gender <- hier_create(root = "Total", nodes = c("male", "female"))
hier_display(dim.gender)

# create a named list with each element being a data-frame
# containing information on one dimensional variable and
# the names referring to variables in the input data
dimList <- list(region = dim.region, gender = dim.gender)

# third column containts a numeric variable
numVarInd <- 3

# no variables holding counts, numeric values, weights or sampling
# weights are available in the input data
# creating an problem instance using numeric indices
p1 <- makeProblem(
  data = microdata1,
  dimList = dimList,
  numVarInd = 3 # third variable in `data`
)

# using variable names is also possible
p2 <- makeProblem(
  data = microdata1,
  dimList = dimList,
  numVarInd = "val"
)

# what do we have?
print(class(p1))

# have a look at the data
df1 <- sdcProb2df(p1, addDups = TRUE,
  addNumVars = TRUE, dimCodes = "original")
df2 <- sdcProb2df(p2, addDups=TRUE,
  addNumVars = TRUE, dimCodes = "original")
print(df1)

identical(df1, df2)

S4 class describing a sdcProblem-object

Description

An object of class sdcProblem contains the entire information that is required to protect the complete table that is given by the dimensional variables. Such an object holds the data itself (slot dataObj), the entire information about the dimensional variables (slot dimInfo), information on all table cells (ID's, bounds, values, anonymization state in slot problemInstance), the indices on the sub tables that need to be considered if one wants to protect primary sensitive cells using a heuristic approach (slot partition and the information on which groups or rather subtables have already been protected while performing a heuristic method (slots startI and startJ).

Details

slot dataObj:

an object of class dataObj (or NULL) holding information on the underlying data

slot dimInfo:

an object of class dimInfo (or NULL) containing information on all dimensional variables

slot problemInstance:

an object of class problemInstance holding information on values, bounds, required protection levels as well as the anonymization state for all table cells

slot partition:

a list object (or NULL) that is typically generated with calc.multiple(type='makePartitions',...) specifying information on the subtables and the necessary order that need to be protected when using a heuristic approach to solve the cell suppression problem

slot startI:

a numeric vector of length 1 defining the group-level of the subtables in which a heuristic algorithm needs to start. All subtables having a group-index less than startI have already been protected

slot startJ:

a numeric vector of length 1 defining the number of the table within the group defined by parameter startI at which a heuristic algorithm needs to start. All tables in the group having an index j smaller than startJ have already been protected

slot indicesDealtWith:

a numeric vector holding indices of table cells that have protected and whose anonymization state must remain fixed

Note

objects of class sdcProblem are typically generated by function makeProblem and are the input of functions primarySuppression and protectTable

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

A Problem-Instance used for examples/testing

Description

sdc_testproblem() returns a sdc-problem instance with 2 hierarchies and optionally with a single suppressed cell that is used in various examples and tests.

Usage

sdc_testproblem(with_supps = FALSE)

Arguments

Value

a problem instance

Examples

p1 <- sdc_testproblem(); p1
sdcProb2df(p1)

# a single protected cell
p2 <- sdc_testproblem(with_supps = TRUE); p2
sdcProb2df(p2)

# cell status differs in one cell
specs <- c(gender = "female", region = c("A"))
cell_info(p1, specs = specs)
cell_info(p2, specs = specs)

modify cutList-objects depending on argument type

Description

modify cutList-objects depending on argument type

Usage

set.cutList(object, type, input)

## S4 method for signature 'cutList,character,list'
set.cutList(object, type, input)

Arguments

• addCompleteConstraint: add a constraint to argument object

• removeCompleteConstraint: remove a constraint from argument object

• type==addCompleteConstraint: a list of length 1

  • first element: an object of class cutList with exactly one constraint

• type==removeCompleteConstraint: a list of length 1

  • first element: numeric vector of length 1 specifying the index of the constraint that should be removed

Value

an object of class cutList

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

modify dataObj-objects depending on argument type

Description

modify dataObj-objects depending on argument type

Usage

set.dataObj(object, type, input)

## S4 method for signature 'dataObj,character,listOrNULL'
set.dataObj(object, type, input)

Arguments

• rawData: set slot 'rawData' of argument object

• type==rawData: a list containing raw data

Value

an object of class dataObj

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

modify dimInfo-objects depending on argument type

Description

modify dimInfo-objects depending on argument type

Usage

set.dimInfo(object, type, input)

## S4 method for signature 'dimInfo,character,character'
set.dimInfo(object, type, input)

Arguments

• strID: set slot 'strID' of argument object

• type==strID: a character vector containing ID's

Value

an object of class dimInfo

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

change linProb-objects depending on argument type

Description

change linProb-objects depending on argument type

Usage

set.linProb(object, type, input)

## S4 method for signature 'linProb,character,list'
set.linProb(object, type, input)

Arguments

• objective: change coefficients of the objective

• direction: change vector of direction of the constraints

• rhs: change vector of right hand side of the constraints

• types: change vector of bounds of the objective variables

• bounds: change bounds of the objective variables

• constraints: change constraint matrix

• removeCompleteConstraint: remove a specific constraint from the object

• addCompleteConstraint: add a constraint to the object

• type==objective: a list of length 1

  • first element: numeric vector defining coefficients of the objective

• type==direction: a list of length 1

  • first element: character vector defining direction of the constraints

• type==rhs: a list of length 1

  • first element: numeric vector defining right hand side of the constraints

• type==types: a list of length 1

  • first element: character vector defining types of objective variables

• type==bounds: a list of length 2

  • element 'lower': a list with the first element containing indices and the second element containing corrsponding lower bounds

  • element 'upper': a list with the first element containing indices and the second element containing corrsponding upper bounds

• type==constraints: a list of length 1

  • first element: an object of class simpleTriplet

• type==removeCompleteConstraint: a list of length 1

  • first element: numeric vector of length 1 defining the index of the constraint that should be removed

• type==addCompleteConstraint: a list of length 1

  • first element: an object of class cutList defining the constraint that should be added

Value

an object of class linProb

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

modify problemInstance-objects depending on argument type

Description

modify problemInstance-objects depending on argument type

Usage

set.problemInstance(object, type, input)

## S4 method for signature 'problemInstance,character,list'
set.problemInstance(object, type, input)

Arguments

• lb: set assumed to be known lower bounds

• ub: set assumed to be upper lower bounds

• LPL: set lower protection levels

• UPL: set upper protection levels

• SPL: set sliding protection levels

• sdcStatus: change anonymization status

• element 'indices': numeric vector defining the indices of the cells that should be modified

• element 'values': numeric vector whose values are going to replace current values for cells defined by 'indices' depending on argument type

Value

an object of class problemInstance

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

modify sdcProblem-objects depending on argument type

Description

modify sdcProblem-objects depending on argument type

Usage

set.sdcProblem(object, type, input)

## S4 method for signature 'sdcProblem,character,list'
set.sdcProblem(object, type, input)

Arguments

• problemInstance: set|modify slot 'problemInstance' of argument object

• partition: set|modify slot 'partition' of argument object

• startI: set|modify slot 'startI' of argument object

• startJ: set|modify slot 'startJ' of argument object

• indicesDealtWith: set|modify slot 'indicesDealtWith' of argument object

• an object of class problemInstance if argument type matches 'problemInstance'

• a list (derived from calc.multiple(type='makePartition', ...) if argument type matches 'partition'

• a numeric vector of length 1 if argument type matches 'startI' or 'startJ'

• a numeric vector if argument type matches 'indicesDealtWith'

Value

an object of class sdcProblem

Note

internal function

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Set/Update information in sdcProblem or problemInstance objects

Description

Function setInfo() is used to update values in sdcProblem or problemInstance objects

Usage

setInfo(object, type, index, input)

Arguments

Value

a sdcProblem- or problemInstance object

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

Examples

# load example-problem with suppressions
# (same as example from ?primarySuppression)
p <- sdc_testproblem(with_supps = TRUE)

# which is the overall total?
idx <- which.max(getInfo(p, "freq")); idx

# we see that the cell with idx = 1 is the overall total and its
# anonymization state of the total can be extracted as follows:
print(getInfo(p, type = "sdcStatus")[idx])

# we want this cell to never be suppressed
p <- setInfo(p, type = "sdcStatus", index = idx, input = "z")

# we can verify this:
print(getInfo(p, type = "sdcStatus")[idx])

# changing slot 'UPL' for all cells
inp <- data.frame(
  strID = getInfo(p, "strID"),
  UPL_old = getInfo(p, "UPL")
)
inp$UPL_new <- inp$UPL_old + 1
p <- setInfo(p, type = "UPL", index = 1:nrow(inp), input = inp$UPL_new)

show objects of class sdcProblem-class.

Description

just calls the corresponding print-method

Usage

## S4 method for signature 'sdcProblem'
show(object)

Arguments

S4 class describing a simpleTriplet-object

Description

Objects of class simpleTriplet define matrices that are stored in a sparse format. Only the row- and column indices and the corresponding values of non-zero cells are stored. Additionally, the dimension of the matrix given by the total number of rows and columns is stored.

Details

slot i:

a numeric vector specifying row-indices with each value being geq 1 and leq of the value in nrRows

slot j:

a numeric vector specifying column-indices with each value being geq 1 and leq of the value in nrCols

slot v:

a numeric vector specifying the values of the matrix in cells specified by the corresponding row- and column indices

slot nrRows:

a numeric vector of length 1 holding the total number of rows of the matrix

slot nrCols:

a numeric vector of length 1 holding the total number of columns of the matrix

Note

objects of class simpleTriplet are input of slot constraints in class linProb-class and slot slot con in class cutList-class

Author(s)

Bernhard Meindl bernhard.meindl@statistik.gv.at

summarize object of class sdcProblem-class or safeObj-class.

Description

extract and show relevant information stored in object ofs class sdcProblem-class or safeObj-class.

Usage

## S4 method for signature 'sdcProblem'
summary(object, ...)

Arguments

Write a problem in jj-format to a file

Description

This function allows to write a problem instance in JJ-Format to a file.

Usage

writeJJFormat(x, tabvar = "freqs", path = "out.jj", overwrite = FALSE)

Arguments

Value

invisibly the path to the file that was created.

Examples

# setup example problem
# microdata
utils::data("microdata1", package = "sdcTable")

# create hierarchies
dims <- list(
  region = sdcHierarchies::hier_create(root = "Total", nodes = LETTERS[1:4]),
  gender = sdcHierarchies::hier_create(root = "Total", nodes = c("male", "female")))

# create a problem instance
p <- makeProblem(
  data = microdata1,
  dimList = dims,
  numVarInd = "val")

# create suitable input for `writeJJFormat`
inp <- createJJFormat(p); inp

# write files to disk
# frequency table by default
writeJJFormat(
  x = inp,
  path = file.path(tempdir(), "prob_freqs.jj"),
  overwrite = TRUE
)

# or using the numeric variable `val` previously specified
writeJJFormat(
  x = inp,
  tabvar = "val",
  path = file.path(tempdir(), "prob_val.jj"),
  overwrite = TRUE
)