HIERLEADSTRING

[ppdewolf]

Please specify
SDC tool used: tau-argus
Version used: 4.2.2 build 1
Operating system used: Windows

---

From eamil exchange with Annu Cabrera:

In a nutshell: from microdata I specify two tables that I wish to protect with the Linked Tables -option. Both tables have hierarchical nace code as one of the explanatory variables. In the original hrc-file the leading string is ‘.’ (dot) and this is correctly specified in the original rda-file. However, when Tau-Argus creates the cover table and associated hrc- and rda-files the error happens. In the new hrc-file a dot still appears as a leading string but in the rda-file for some reason the leading string is specified as ‘@’. Due to this error Tau-Argus cannot recognize the nace codes in the cover table correctly and the protection process is stopped.

Attached to this message (HEARLEADSTRING.zip) are the original microdata file (fake data) with associated rda- and hrc-files. I also attached a picture of the error message, a logbook file and the cover table files produced by Tau-Argus (I just copied them from my temp directory).

As I said, I made two tables. First one is nace X expl1 with resp as the response variable. The second table is nace X expl1 X expl2 with resp as the response. In the first table nace is at 2-digit level and in the second table at 1-digit level. There are some extra sub-totals in the hierarchy file, i.e. not just the standard nace classification. In the microdata file the nace codes are at 5-digit level so I used the ‘Recode’ button in Tau.

I use Tau-Argus version 4.2.2 build 1.

Do you have any idea what might cause this error? Of course easy solution (that I actually used) is to use ‘@’ as a leading string in the original hrc-file. But one might have reasons to use some other character and then this easy solution is not that convenient.

[ppdewolf]

I can reproduce the error, so thanks for the example files.
Playing a bit with the example, I got even stranger results: it only happens when both tables have a (from within argus) recoded NACE classification. If I recode only one of the tables, the HIERLEADSTRING is correctly a "." in the rda-file of the cover table.

Also got me wondering what will happen when the hrc-file of table 1 and the one of table 2 have different HIERLEADSTRINGS to start with and we then recode both.

I'll look into it more closely at some time ;-)