It seems that groupVars deals with integers in parameter before and after.
Another strategy of dealing with the levels should be defined since this runs without warning, but produces wrong results:

require(sdcMicro); data(testdata, package="sdcMicro")
sdc <- createSdcObj(testdata,
          keyVars=c('urbrur','water','sex','age'), 
          numVars=c('expend','income','savings'),
          pramVars=c("walls"), 
          w='sampling_weight', 
          hhId='ori_hid')

labs <- c("1-9","10-19","20-29","30-39",
          "40-49","50-59","60-69","70-79","80-130")
sdc <- globalRecode(sdc, column="age",
                    breaks=c(0,9,19,29,39,49,59,69,79,130), 
                    labels=labs)

table(extractManipData(sdc)$age)

1-9 10-19 20-29 30-39 40-49 50-59 60-69 70-79 80-130
1128 1110 635 629 447 250 200 70 13

sdc <- groupVars(sdc, var="age",
                    before=labs, 
                    after=c("1-9","10-19","20-29","30-39",
                             "40-49","50-59","60-69","70-130"))
table(extractManipData(sdc)$age)

1-9 10-19 20-29 30-39 40-49 50-59 60-69 70-130
1141 1110 635 629 447 250 200 70

Based on Joerg Drechslers e-mail to us, reporting a problem with mdav.

microaggregation gives wrong results for method mdav (old slow version and IHSN version) for the following toy data:

library(sdcMicro)
toyData <- data.frame(x=c(2,3,3,20,21), y=c(1,2,2,19,20))
toyData
   x  y
1  2  1
2  3  2
3  3  2
4 20 19
5 21 20

The result should be:

1  2.666667  1.666667
2  2.666667  1.666667
3  2.666667  1.666667
4  20.500000 19.500000
5  20.500000 19.500000

However, the outcome is:

m3 <- microaggregation(toyData, method="mdav", aggr=2)
m3$mx ## wrong!
m3$mx
     x    y
1  2.5  1.5
2  3.0  2.0
3  2.5  1.5
4 20.5 19.5
5 20.5 19.5

wrong.

By the way, the old version gives the correct values but the ordering of observations is not correct:

m3 <- microaggregation(toyData, method="mdavold", aggr=2)
m3$mx ## wrong!
m3$mx
        [,1]      [,2]
3   2.666667  1.666667
4  20.500000 19.500000
5  20.500000 19.500000
31  2.666667  1.666667
41  2.666667  1.666667

GUI-Guidelines (from menu, Help): include latest version of guidelines

dat <- data.frame(x=rnorm(100))
microaggregation(obj=dat, variables="x", method="mdav", aggr=3, measure="mean")

-> error

dat <- data.frame(x=rnorm(100), y=rnorm(100))
microaggregation(obj=dat, variables="x", method="mdav", aggr=3, measure="mean")
-> works

microaggregation(obj=dat, variables=c("x","y"), method="mdav", aggr=3, measure="mean")
-> works but inconsistent output (invisible())

The plots in the Recode window, even for very small examples are not there.

Frequencies (age, after recoding the age, e.g. the frequencies of the old
age classes are also displayed as xx.1, xx.2, .... --> correct this)

require(sdcMicro); data(testdata)
sdc <- createSdcObj(testdata,
keyVars=c('urbrur','water','sex','age'),
numVars=c('expend','income','savings'),
w='sampling_weight', hhId='ori_hid')
length(table(sdc@manipKeyVars[,"age"])) #OK
p1 <- print(sdc)
sdc <- globalRecode(sdc, column="age",
breaks=c(1,9,19,29,39,49,59,69,100), labels=1:8)
length(table(sdc@manipKeyVars[,"age"])) #OK
p2 <- print(sdc)

identical(p1,p2)

TRUE

Risk", "Protection", "Information Loss" (in Tab Categorical, Continuous)
larger or bold to have it more visible.

since there is now no function pram_strata(), we have to options to deal with this situation in the docs.

1. remove all references to pram_strata() with correct code
2. add a wrapper function pram_strata() that just calls pram()

the same has to be done for references to the "old" pram function, since parameters have been changed.

thoughts?

require(sdcMicro); data(testdata)
sdc <- createSdcObj(testdata,
keyVars=c('urbrur','water','sex','age'),
numVars=c('expend','income','savings'),
#pramVars=c("walls"),
w='sampling_weight',
hhId='ori_hid')

sdc <- pram(sdc)

error

sdc <- createSdcObj(testdata,
keyVars=c('urbrur','water','sex','age'),
numVars=c('expend','income','savings'),
pramVars=c("walls"),
w='sampling_weight',
hhId='ori_hid')

sdc <- pram(sdc)

error

Help menu:

• R sdcMicro Index
• R sdcMicro Help Files

it should be possible to undo a recode to factor if user were specifying breaks

Seems output=input ?!

explain: expected number of re-identifications !!! (in the GUI Tutorial)

UNDO: document somewhere that x undo's are possible for data size of x: >100000 kein UNDO, sonst ist maxUndo auf 1 gesetzt ist aber beliebig änderbar (maybe only in the GUI tutorial)

problem: es koennen keine "nicht-keyVars" als variablen fuer pram verwendet werden. gleichzeitig ist es nicht möglich, localSuppression() nur auf ein subset von "keyVars" auszuführen. Ich denke, folgendes sollte möglich sein:

• data(free1)
• sdc <- createSdcObj(free1, keyVars=c("REGION","SEX","AGE","MARSTAT","KINDPERS"),weightVar="WEIGHT")
• sdc <- localSuppression(sdc, vars=c("REGION","SEX","AGE", k=3)
• sdc <- pram(sdc, vars=c("MARSTAT","KINDPERS")
• sdc <- pram(sdc, vars=c("EDUC1")

zum programmieren in sdcMicro scheint es recht einfach zu sein, aber ich bin mir nicht sicher, ob das nicht auswirkungen auf das gui und andere sachen (extractManipData(), get.sdcMicroObj(..., type="manipData")) hat. alex?

would be nice to have. print function should show basic things about the object
(data size, key variables)

Change it?!?

in Window "Recode" now "rename selected level" and "group selected levels" instead of "rename" and "group", war das gemeint?

in general: abbrechen --> cancel. Can that be changed even one uses German
Language. Mix of Languages is not good...

OK and cancel in the same order. in some windows Cancel is left to OK, in
some it is right to OK. -- Suchen, wo!?!?

• confirm that the direct identifiers are removed in tab identifiers. Move
  "remove direct identifiers" to variable selection frame! and instead of
  "remove direct identifiers" do this confirmation button. Eigentlich muesste
  man das fuer den Report auch beruecksichtigen....

change the sentence "Reported is the number | ..." --> should be better
readable and understandable.

I have installed all required packages for sdcMicroGUI and it seems that it works OK but when i try to select variables to be categorical, numerical, weight etc and when i click OK windows appear and it shows script running but it never finishes. and in the R console appears
Error in measure_riskWORK(manipData, keyVars, w = w, hid = hhId, ...) :
Please define valid key variables
What is the problem?
Thanks

we use "data(ses)" in a vignette --> add "laeken" to deps?

Critical Revision: Make an UNDO also possible in the Recode window
(example: one recodes age in 6 classes, but then he want to reject this and
want to have 7 classes)

wenn die Variable anschließend als stratavar im sdcObject ausgewählt ist.

==> Create_strata_var muss ins Skript

sometimes the individual risks evaluate to NaN. In this case localSupp() threw an error. We can easily work around this issue but we have to look at the problems with the individual risk calculation.

only when using a sdcMicroObj, otherwise it works.

data(testdata)
sdc <- createSdcObj(testdata,
keyVars=c('urbrur','roof','walls','water','electcon','relat','sex'),
numVars=c('expend','income','savings'), w='sampling_weight')
sdc <- mafast(sdc,by="urbrur")

mafast finds no keyVars for applying mafast on strata.

4000 might be too low now with the quiet fast system, maybe set it to 20000?

The new warning text is: Large data sets require extensive computation time, so please be patient.

short explanation what means fk and Fk( with overlay?) -- tooltip added fk=sample frequency\nFk=grossed up population frequency

library(sdcMicro)
[snip]
library(far)
Loading required package: nlme
far library : Modelization for Functional AutoRegressive processes
version 0.6-0 (2005-01-10)
data(Tarragona)
a1 <- addNoise(Tarragona)
a2 <- addNoise(Tarragona,method="ROMM")
[1] "please load package far"

[[it freezes here and says i have not loaded far even though i have]]
[[i just kill the process otherwise it overloads my computer]]

sessionInfo()
R version 3.1.0 (2014-04-10)
Platform: i386-w64-mingw32/i386 (32-bit)

locale:
[1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United States.1252
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252

attached base packages:
[1] stats graphics grDevices utils datasets methods base

other attached packages:
[1] far_0.6-3 nlme_3.1-117 sdcMicro_4.3.0 xtable_1.7-3 data.table_1.9.2
[6] knitr_1.6 brew_1.0-6

loaded via a namespace (and not attached):
[1] car_2.0-20 class_7.3-10 cluster_1.15.2 DEoptimR_1.0-1 e1071_1.6-3
[6] evaluate_0.5.3 formatR_0.10 grid_3.1.0 lattice_0.20-29 MASS_7.3-31
[11] nnet_7.3-8 plyr_1.8.1 Rcpp_0.11.1 reshape2_1.2.2 robustbase_0.91-1
[16] sets_1.0-13 stringr_0.6.2 tools_3.1.0

I just installed XQuartz and GTK+ on OS X 10.9.1 but get the following error message:
( I restarted R of course.)

The GTK+ installation screen keeps prompting.

If the package still does not load, please ensure that GTK+ is installed and that it is on your PATH environment variable
IN ANY CASE, RESTART R BEFORE TRYING TO LOAD THE PACKAGE AGAIN
Loading required package: gWidgets
Error : .onAttach failed in attachNamespace() for 'gWidgetsRGtk2', details:
  call: .Call(name, ..., PACKAGE = PACKAGE)
  error: "S_gtk_icon_factory_new" not available for .Call() for package "RGtk2"
In addition: Warning message:
Failed to load RGtk2 dynamic library, attempting to install it. 
Error: package ‘gWidgetsRGtk2’ could not be loaded

ohne das jetzt ueberpruefen zu koennen - ich glaube der print output unterscheidet sich bei den beiden Funktionen, wobei pram() detailiertere Info gibt, zumindest falls nur eine Variable gepramt wird.

set.seed( 657613 )
activedataset <- testdata
sdcObject <- createSdcObj(activedataset,keyVars=c('urbrur','roof','walls'),numVars=c('income'),weightVar=c('sampling_weight'),hhId=c('ori_hid'))
sdcObject <- microaggregation(sdcObject,aggr=c( 3 ), method=c('mdav'), variables=c('income'))
sdcObject <- microaggregation(sdcObject,aggr=c( 3 ), method=c('mdav'), variables=c('income'),strata_variables=c('urbrur'))
sdcObject@risk$numeric
s=dRisk(sdcObject)
s@risk$numeric

after selection of variables the one window which pops up - change text
there: something like "...but not recognized as correct format. This can be
confirmed or changed in the next window." --fixed

due to update of data.table, localSuppressionWORK does not work and thus a test ("tests/reporting_test.R") fails.

No idea what that means, since there is no "type" in the Recode window in the frame "Recode to factor"