SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

With tau-Argus 3.5.0 you could specify a cost and a shadow variable when creating a table from micro data. But with 4.1.3 you get an error if you have or in the meta data file. Is this a bug? In the manual this option is indeed only for tabulated data.

Lars-Erik Almberg Staistics Sweden

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows
Hi!

Hypercube stops working in case when explanatory variables are too long.

Best regards,
Andreja.

SDC tool used: Tau-Argus
Version used: 4.1.4
Operating system used: Windows

problem

when creating a batch-file, some options (e.g Apply weights and Missing==safe) in Specify ->Tables seem to be dropped when the batch file is written.

steps to reproduce

• use the following microdata
• use the following metadata file
• File -> Open Microdata -> select the above data and metadata files
• Specify -> Tables
• create first table
  • select region and gender as explanatory variables
  • select num1 as response variable
  • use a minimum frequency FREQ(5,20) and DONOTcheck either Apply weights orMissing==safe
  • add this as the first table
• create second table table
  • use same settings as above but DO check Apply weights andMissing==safe
  • add this as the secondtable

In the overview table at the bottom, the two table seem to be different as they have different values in the column Shadow & cost var

• press compute tables
• Use Output -> Write Batch File to create a Batch File. This file now has the following content (showing only the relevant sections):

<OPENMICRODATA>    "microdata.asc"
<OPENMETADATA>     "metadata.rda"
<SPECIFYTABLE>     "region""gender"|"num1"||
<SAFETYRULE>       FREQ(5,20)
<SPECIFYTABLE>     "region""gender"|"num1"||
<SAFETYRULE>       FREQ(5,20)
<READMICRODATA>
<GOINTERACTIVE>

the two tables do not differ in any way.

Version used: 4.1.3

Windows

Bij controlled rounding: partitioning geeft error. Lijkt een licensie error te zijn?

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

Apply an apriori-file to a table and save the resulting table. The used apriori file will be stated in the report file. When you then recompute the same table and apply another apriori-file and save the resulting table, the first apriori-file info is still written to the report file, i.e., the "old" (no longer valid) apriori-file info is written to the report file. Even though the correct (second) apriori-file is applied to the table.

Navigation within the file dialog windows would be more comfortable if one could define a working directory under the menu TauArgus options.

SDC tool used: tau-argus
Version used: 4.1.3-build 1
Operating system used: Windows

SDC tool used: tau-argus, sdcTable
Version used: 4.1.4
Operating system used: Windows

Reading non-hierarchical (additive or non-additive) tabular input, selecting "Allow non-additivity", suppress with "Modular", results in error "Error in modular suppression procedure HiTaS: complete table has to be skipped?".

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

Is it possible to get top1 and top2 at holdinglevel in the intermediate output after primary suppression when you run it with holding? It would be very useful for us when we work with consent control. Now top1 and top2, in the intermediate output, are the the two largest contributions not including holding information. It´s no bug, holding works as it should.
Maybe there is a reason for reporting top1 and top2 in the way it´s done now?

Lars-Erik Almberg Statistics Sweden

Version used: 4.1.3
Windows
Fixed

example data file.txt

SDC tool used: Tau-Argus
Version used: 4.1.3 Build 1
Operating system used: Windows

When trying to suppress a table from a .tab file, the Modular method crashes if the values of explanatory variables are numbers, containing a decimal period. (A common example are NACE codes.) Meanwhile, Hypercube and Optimal work as expected.

Eliminating the dots from the table and the hierarchy file resolves this issue.

Here is a MWE:
asc

TOT;4319;114;S;1;1
85;4140;74;S;1;1
85.1;2592;18;S;1;1
85.2;648;12;S;1;1
85.3;;0;S;1;0
85.4;864;8;S;1;1
85.5;36;36;S;1;1
86;;0;S;1;0
86.9;;0;S;1;0
88;;0;S;1;0
88.1;;0;S;1;0
90;149;20;U;13.5;13.5
90.0;149;20;U;13.5;13.5
93;30;20;S;1;1
93.1;30;20;S;1;1

rda

<SEPARATOR>         ";"                                                                      
NACE
                    <RECODEABLE>                                                             
                    <TOTCODE>           TOT                                                  
                    <HIERARCHICAL>                                                           
                    <HIERCODELIST>      "NACEshort.hrc"
                    <HIERLEADSTRING>    "@"                                                  
response                                                                                     
                    <NUMERIC>                                                                
freq                                                                                         
                    <FREQUENCY>                                                              
status                                                                                       
                    <STATUS>                                                                 
lpl                                                                                          
                    <NUMERIC>           <LOWERPL>                                            
upl                                                                                          
                    <NUMERIC>           <UPPERPL>                                            

   

NACEshort.hrc

85
@85.1
@85.2
@85.3
@85.4
@85.5
86
@86.9
88
@88.1
90
@90.0
93
@93.1

SDC tool used: tau-argus
Version used: 4.1.4
Operating system used: Windows 7

Hello,

Tau-Argus is wrongly stating that a table (*.tab) is not additive, while it obviously is.

Apparently, the problem occurs when the response variable has decimals (the same data converted to integers works).

Furthermore, TA is unable to point out which cells cause the problem. The information text is empty.

There follows a simple example that should reproduce the bug (I attached the files as well if you need to test : additive_tables.zip).
I checked this is actually additive. Furthermore, version 3.5.0 does not complain about non-additivity.

decimals1.tab

A1;B1;70911.7;40
A1;B2;64512.5;447
A1;TOTAL;135424.2;451
A2;B1;2856.1;5
A2;B2;2276;65
A2;TOTAL;5132.1;65
TOTAL;B1;73767.8;40
TOTAL;B2;66788.5;447
TOTAL;TOTAL;140556.3;451

decimals1.rda

<SEPARATOR> ;
<SAFE> S
<UNSAFE> U
<PROTECT> P
V1
  <RECODEABLE>
  <TOTCODE> TOTAL
V2
  <RECODEABLE>
  <TOTCODE> TOTAL
RESP
  <NUMERIC>
  <DECIMALS> 1
TOP1
  <MAXSCORE>

This version with integer values for response variable works both on 3.5.0 and open source version.

decimals0.tab

A1;B1;70912;40
A1;B2;64513;447
A1;TOTAL;135425;451
A2;B1;2856;5
A2;B2;2276;65
A2;TOTAL;5132;65
TOTAL;B1;73768;40
TOTAL;B2;66789;447
TOTAL;TOTAL;140557;451

decimals0.rda

<SEPARATOR> ;
<SAFE> S
<UNSAFE> U
<PROTECT> P
V1
  <RECODEABLE>
  <TOTCODE> TOTAL
V2
  <RECODEABLE>
  <TOTCODE> TOTAL
RESP
  <NUMERIC>
TOP1
  <MAXSCORE>

Can you confirm this is a bug ? Let me know if you need more information.

Regards,

Pierre-Yves

SDC tool used: mu-argus
Version used: 5.1.1
Operating system used: Windows

Trying to use synthetic data option (hybridIPSO) from mu-argus. Get error can't find file specified.

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

We have tried the new version at Statistics Sweden and found some bugs.

Interface:

1. You can´t have Swedish letters (å,ä and ö) in the path to the folder were the files are located.
2. Specify metadata. If you choose free format the new variables are not saved. You have to click OK and leave. Then if you reenter Specify metadata it works as it should.
3. Specify tables. If you choose p% the minimum frequency is automaticly included, even if you don´t choose minimum frequency. You notice that when you write a batchfile.
4. Specify tables. If you choose Zero unsafe tau-Argus include a decimal in the range. Suddenly 10 (for example) becomes 10.0 and then you can´t build a table. If you write a batch file the parameter will be 10.0 and you can´t Open batch process without error.
5. We also had problems with decimals when reloading a table back into tau-Argus for linked secondary suppression. Tau-Argus said the tables (that it has created) were non additative. I removed the decimals and then it was no problem.

Batch mode:

1. In batch mode tau-Argus don´t produce a log file into the specified folder.

Best Regards,
Lars-Erik Almberg Statistics Sweden

Dear, Tau argus team,

At beginning I read micro and meta data using tau argus 4.1.2 build 1. When specified table I get error, that tau argus are closing and do not make any calculations.

In attachment I add print screen how look window before select compute table and java error.

Maybe you can explain why it happens?

With best wishes,

Ilona Kallione

Version used: 4.1.3

Windows

In interactive mode kun je overlappende variabelen inlezen. Je krijgtwel een "warning", maar je kunt wel door gaan. In batch mode leidt het tot een error en stopt tau.
VB: (in rda-file)
SBI 58 4

"Total"

1 2 1 0 0

SBI_long 58 5

"Total"

1 2 1 1 0
Voorstel Georges:
korte termijn: vervang error door warning in batch mode
lange termijn: recoding in batch mode toevoegen (bij gebruik microdata), bv (Tablenumber, "ExpVar", Maxlevel, "GlobalRecodeFilename") en mogelijk ook aan rda-file.

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

Specify metadata. If you choose free format the new variables are not saved. You have to click OK and leave. Then if you reenter Specify metadata it works as it should.

Lars-Erik Almberg, Statistics Sweden

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

We can´t get linked secondary suppression in batch mode to work. The syntax in the arb-file that works with 3.5.0 doesn´t work here (see arb-file in zip file).

Lars-Erik Almberg, Statistics Sweden

Synthetic data:

Linked sec suppression in batch mode.zip

SDC tool used: tau-Argus
Version used: 4.1.3 build 1
Operating system used: Windows 7 Pro

When repeatedly calling tau-Argus in command line using a batch file ( .bat-file) supressed output file comes out empty occationally. There has been cases (larger hierarcy tables) where csv is only partly written. See example tautest.zip attached (check path of TauArgus.exe & test files before run): When batch.bat is run some of supressed output files (..out_5_1.csv, not included) are empty (size of 0 bytes). Sometimes csv is written ok but rda comes out empty. If csv fails rda is missing. When failed commands are run separately everything works.

tautest.zip

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

Open free format file. Protect table. Consecutively open in same session fixed format file. I get an error in reading the microdata.

SDC tool used: Tau-Argus
Version used: 3.5.0 build 27
Operating system used: Windows

My colleague and I got some trouble with hierarchy file attached in pull requests (what is the best option to post an issue?).
It is not possible to use the Hypercube method taking into account this hierarchy file.

We have noticed something weird: it's working if hierarchical variable contains 11 or 13 digits, but it's not working anymore when it contains 12 digits (like in toy microdata attached in pull request).

This problem arises also with Tau-Argus version 4.1 with some real French data, but for now we didn't manage to replicate it with toy data.

SDC tool used: Tau-Argus
Version used: 4.1.4
Operating system used: Windows

When a cell is selected and then "Set to protected" or unsafe, it is deselected. To set another cell to protected or safe, a user must click on it and then go back to "Set to protected" (or unsafe). Going back and forth with a mouse pointer is unavoidable in this scenario.

In earlier versions of Tau-Argus (e.g. 3.5.0), a user could set one cell to protected or unsafe, which stayed selected, and then move to another by using the arrow buttons of the keyboard, without resorting to clicking. The mouse pointer could then rest within the "Change status" panel, which made for quicker operations.

Is there any possibility of bringing this functionality back?

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

In batch mode tau-Argus don´t produce a log file into the specified folder.
We specify the log file like this:
"C:...\SAStmp\tab_1.LOG"

Lars-Erik Almberg Statistics Sweden

SDC tool used: Tau-Argus
Version used: 4.1.4
Operating system used: Windows

When using Hypercube method of protection, lower and upper protection levels are disregarded. Only the (manual) range specified with the safety rules is taken into account. The specifications of protection levels are sensible with other methods of protection (e.g. Modular), however.

It would probably be helpful to reiterate this in the Reference section of the Manual. One possible place for it would be under 4.3.2 File | Open table (p. 79), as this is where the structure of the .tab files is explained. Another possibility would be under 4.4.3 Specify | Metafile [for tabular data] (p. 87) as this is where the fields Lower and Upper protection level are mentioned in the context of an .rda file.

Version used: 4.1.3
Operating system used: Windows

Minimum frequency rule moved to table summary list, even when not selected. Should check if it is also used when not selected.

A first step would be a help file. A more advanced approach would be a context specific ? button, which in each window highlights the purpose of the dialog and each of its features.

The fact that there is a manual is already a great achievement. However, I believe that the feature specific parts of the manual should be moved or at least be duplicated in the software. This will render the tool more accessible.

SDC tool used: tau-argus
Version used: 4.1.3 Build 1
Operating system used: Windows

SDC tool used: tau-argus / mu-argus
Version used: 4.1.x / 5.1.x
Operating system used: Windows

During long running read/write/apply-processes (like write output, read apriori-files) show progress or at least show that it is "busy". Currently you have to wait for a pop-up stating that it has finished, but when you don't know that...

Please specify
SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

When saving a table as "csv"-file, at the end of each row a spurious comma is added. Some systems can't handle that correctly when importing that csv-file.

Version used: 4.1.3

Windows

Hypercube geeft message over "frozen/protected cells". Is verwarrend en soms ook "leeg". Beter achterwege laten?

SDC tool used tau-argus
Version used: 4.1.3 // 4.1.1
Operating system used: Windows

Hello,
I'm Julien Lemasson, I work with Maxime Bergeat at Insee (France).
I have an issue with modular (free solver) of the 4.1.3 version of Tau-Argus.
I tabulate two 3-dimensional tables : 2 hierarchical variables in common + 2 other variables (one in each table)
Number of cells
157_53_5 = 41 605 for table 1
157_53_3 = 24 963 for table 2.

The response variable has no decimals. I use the holding option (but I try without and the issue is the same).
I do linked tables and then launch modular.

With Tau-Argus 4.1.1, that works. It is pretty long (2/3 hours).
With Tau-Argus 4.1.3 (and 4.1.2 too) that crashed pretty fast (1 minute) and a window tells me that :

One of the difference on the temporary files is on the CSPlogfile.txt (I join bothCSP log files)
CSPlogfile - 4.1.3.txt
CSPlogfile - 4.1.1.txt

I don't know what is

ERROR: unknown LP problem status = 1101
ERROR: pstat=1101
but I guess it is related to the "max time per substable" option. I tried few values :
"1" => does not work
"2" and more => does not work
"0" => work well on 4.1.3, and comparatively fast (20 minutes)... but the results are not satisfactory, because Hypercube does better here (according to the response value of secondary cases), in less time (44 seconds).
I wish I could try 10 seconds, or whatever value betwen 0 and 1 minutes, but i don't succeed in using this option, and I don't think it can be.

I try other configurations.

• applicate modular (and optimal too) on only one of the two tables (with 3 dimensions so, including 2 hierarchical), but there, the soft just crash, and the window "error runing the cover table" doesn't even appear.
• applicate modular on one big table with the 4 variables. That does not work at all, the soft just crash in the same lapses of above test.
• applicate modular on the 2 tables, removing the hierarchy on 1 out of 2 hierarchical variable, making the number of cells from 66 568 to 157_48_5+157_48_3= 60 288. That works, it takes 20 minutes.
• applicate modular on the 2 tables, removing the other hierarchy on 1 out of 2 hierarchical variable, making the number of cells from 66 568 to 113_53_5+113_53_3= 47 912. That works, but it seems to be long, I turn off the soft, cause of too long time more than 30 minutes. It uses to crash way more early when it does, so i guess it works.
  These last 2 tests lead me to think that it is not a problem of hrc files in particular, but their accumulation.
• I also change the order of the explanatory variables, but it changes anything.

Anyway, modular uses to work, on others tabulations, but there, on this particular configuration, there is an issue.

If anyone get an idea of what I did wrong (if I did), or how about work around this problem, I am aware !

Best regards :^),
Julien

Please specify
SDC tool used: Tau-Argus
Version used: 3.5.0_build_27 and 4.1.0_build_2
Operating system used: Windows

Instance_2 shows the case, where a secondary suppression is set into a marginal cell instead of the remaining inner cell. We use a lamda value of 0.4 and that is why inner cell and marginal cell obtain the same costs. But also in case of same costs inner cells should be preferred as secondary suppression compared to marginal cells. So there should be added some kind of exta protection for marginal cells.

In case of the littlte instance results are the same for TauArgus V_3.5.0_build_27 and v_4.1.0_build_2 with settings Modular, "Do singletons" and lambda=0.4, but in general we experience this behaviour much more often with version 4.1.

Instance_2.zip

The file dialog windows used in the application do not allow path pasting, forcing a point & click experience for navigation.

SDC tool used: tau-argus
Version used: 4.1.3-build 1
Operating system used: Windows

Version used: 4.3.1
Windows
"Zero unsafe" not used!

SDC tool used: tau-argus
Version used: 4.1.4
Operating system used: Windows

Saving a table as "intermediate", the name of the codelist-file does not contain directory name, hierarchy-file does have directory name in the .rda-file. E.g., (excerpt):

Region "99"
    <RECODEABLE>
    <TOTCODE> "Total"
    <DISTANCE> 2 4 4 4 4
    <CODELIST> "REGION.CDL"
    <HIERARCHICAL>
    <HIERCODELIST> "E:\Users\Peter-Paul\MyDocuments\data\region2.hrc"
    <HIERLEADSTRING> "@"

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

We had problems with decimals when reloading a table back into tau-Argus for linked secondary suppression. Tau-Argus said the tables (that it has created) were non additative. I removed the decimals and then it was no problem.
I include two synthetic micro datasets. _dec is with one decimal. _prim is already primary suppressed datasets that is saved in tau-Argus. I can´t get the _dec-version to work.
Decimals in tabledata.zip

Lars-Erik Almberg Statistics Sweden

Version used: 4.1.3

Windows

After an "error" (for instance when not rounded correctly) then in some cases only zero values in all cells.
After a correct result, and then en 'error', in some cases the old solution returns.

In the table view, it would be practical if an input box was available above the Total column label to allow for (interactive) searching, highlighting or filtering those lines which match the search string provided by the user.

By default, all table levels are collapsed to the lowest level. Clicking on + allows to go one level deeper and so on. Depending on the table complexity, finding any given cell is not necessarily comfortable.

SDC tool used: tau-argus
Version used: 4.1.3-build 1
Operating system used: Windows

SDC tool used: Tau-Argus
Version used: 4.1.4
Operating system used: Windows

To recode a hierarchical variable a user can manipulate a tree structure. In Tau-Argus v. 3.5.0 it was possible to save such a tree structure in a .grc file. This enabled applying the same global recode to many subsequent tables by simply reading the file every time.

In Tau-Argus v. 4.1.4 (and also earlier) this is no longer possible. This makes selecting a global recode more tedious as clicking through the tree structure is necessary every time. The maximum level box may or may not help, depending on the desired recode.

I was wondering about the reason to eliminate this feature and whether it would be possible to reintroduce it.

SDC tool used: tau-argus
Version used: 4.1.4
Operating system used: Windows

Tau-Argus crashes when you specify "Apply weights" and then "Compute tables". This worked with 4.1.3.

Weight.zip

Lars-Erik Almberg Staistics Sweden

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

Specify tables. If you choose Zero unsafe tau-Argus include a decimal in the range. Suddenly 10 (for example) becomes 10.0 and then you can´t build a table. If you write a batch file the parameter will be 10.0 and you can´t Open batch process without error.
To be more specified this don´t happen the first time you build the table. But if you reverse with arrow up to try some other settings the 10 becomes 10.0 and you can´t build the table with arrow down.
I include an example with synthetic micro data.

Lars-Erik Almberg Statistics Sweden
Decimals zero unsafe.zip

SDC tool used: tau-argus
Version used: 4.1.3
Operating system used: Windows

When we do consent control at Statistics Sweden we first run SAS2Argus (tau-Argus) for primary suppression. Then we do consent control and write an apriori file with the cells that should be set to safe. When we run SAS2Argus the second time to do secondary suppression we use the tabulated result from the first run because tau-Argus don't understand the code 'Total' in the apriori file if we use the micro data option.

This is a good solution exept in the case where there are negative contributions to a cell. In this case it might happen that top1 (in absolut value) is larger than the sum of the cell and then tau-Argus will not accept the table. This can be solved by taking the absolut value on every contribution in micro data and do some manually inspection of the result.

But the best way to do this would be to use microdata in the second run combined with the apriori file. I include some made-up data where it is possible to set the 'Total' cell to safe with the apriori function if you load it as table data but it´s not possible to do this if you load it as micro data. This is no bug I guess, it was the same with 3.5.0, but maybe this colud implemented in newer versions of tau-Argus?

Lars-Erik Almberg, Statistics Sweden

Apriori.zip

SDC tool used: mu-argus
Version used: 5.1.1
Operating system used: Windows

When specifying the metadata for "free format" and/or "free format with meta" the separator is not saved in the rda-file. When again using that rda-file, it is not recognized to be free format any more.

SDC tool used: Tau-Argus
Version used: 4.1.4
Operating system used: Windows

When saving a protected table (code-value format), the variable names are not written to the output-file even if Variable names on first row has been selected. The same happens, if I specify e.g.

<WRITETABLE> (1, 3, AS+SE-FL+, "somefile.txt")

in a batch-input file.

Please specify
SDC tool used: Tau-Argus
Version used: 3.5.0_build_27 and 4.1.0_build_2
Operating system used: Windows

Instance_1 shows a case, where a secondary suppression gets an internal safety range, even though "manual safety range" is set to 0% in "Specify table". In the example the cell value of the corresponding primary suppressed cell is not large enough and because if this a third cell (in this case the marginal cell) is suppressed by Tau Argus Modular.

The corresponding cells of the example are C and C1 (Var_1) cross combined with 01 and Total (Var_2). In case of a primary suppression in cells (C101) and (C01), secondary suppressions of (C1Total) and (CTotal) are avoided. Because of this it would be very helpful if the safety range for secondary suppressions is controllable even if some underprotection (across subtables) has to be taken into account.

Behaviour is the same for TauArgus V_3.5.0_build_27 and v_4.1.0_build_2 with settings Modular and "Do singletons".

Instance_1.zip

Version used: 4.3.1

Windows

C,6,S
C1,,S
C2,,S
C3,4,S
C4,2,U
C5,,S
Geeft bij inlezen foutmelding dat niet additief is (6 <> 1 + 1 + 4 + 2 + 1 = 9)

TWEE DEELPROBLEMEN:
2016004_A In een tab-file verbieden dat een variabele leeg (of nul) is met een status ongelijk aan E
2016004_B Vreemde protection levels bij unsafe (U) cellen als variabele (ook) als response variabele wordt gebruikt

SDC tool used: TauArgus
Version used: Version 4.1.1 build 1
Operating system used: Windows

Hi,

When saving a protected table (.sbs format) in a repository with an accented character, only the .html report is saved (the . sbs table is not saved). There is no error message warning the user that the table is not saved properly (it "looks like" everything worked well).

Best regards,
Maxime

Version used: 4.1.3

Windows

In de Generate Apriori kan niet langer de costs als enige worden gekozen. Er moet een status gekozen worden. "Omit" zorgt er voor dat de cel helemaal niet wordt weggeschreven.

Fixed

Version used: 4.1.3
Operating system used: Windows

Level number in table-view differs from number in recode-tree. Max 1 level in recode tree = max 2 level in view table.

In addition, the contents of the file will go unmodified.

SDC tool used: tau-Argus
Version used: 4.1.3-build 1
Operating system used: Windows

Version used: 4.1.3

Windows

Scaled costs (lambda=0.4) lijkt niet altijd te werken zoals verwacht.
Als je de costs scaled, komt er bij modular en optimal iets anders uit (één dimensionaal voorbeeld).
Wellicht: optimal = eerst schalen naar 20000 dan integer weight, modular = eerst integer weight dan schalen naar 20000?

SDC tool used: Tau-Argus
Version used: 4.1.4
Operating system used: Windows

Trying to use this simple batchfile, I get the following error:

Error in batch file
Command: <READTABLE>
This keyword (<READTABLE>) is not allowed in this position

The input data are correct because I can upload (and protect) them successfully using File -> Open Table.