Lang query through RDF_BOX

Virtuoso Universal Server Multi-Model Data Management
,
1

Hello dear Virtuoso community,

Does anyone know how to list all languages, used in the system through RDF_BOX?

We tried the following query:

select DISTINCT rdf_box_lang (__ro2sq(O)) from RDF_QUAD;

But only got default language, 257.

Although we use the following languages in our database.

en 258
ru 259
all 260
de 261
262

This question is triggered by error:

GPF: Dkbox.c:1616 rb lang out of range

Maybe someone had the same?

https://github.com/openlink/virtuoso-opensource/issues/882

We tried to draw main tables here.

Here is our test database if someone could help (login: dba, password: dba).

Thank you!
Stas

2

Dear Virtuoso,

We might have found a heavy bug on our side.

Since we have introduced new language tag @all, we have changed this source code, that checks language tags.

create function DB.DBA.rdf_strlang_impl (in str varchar, in lang any)
{

lang := cast (lang as varchar);
if ((lang is null) or (regexp_match ('^(([a-z][a-z](-[A-Z][A-Z])?)|(x-[A-Za-z0-9]+))\044', lang) is null))
signal ('22007', 'Function rdf_strlang_impl needs a valid language ID as its second argument');
return DB.DBA.RDF_MAKE_LONG_OF_TYPEDSQLVAL (str, null, cast (lang as varchar));
}
;

We have removed regular expression checking.
It became:

create function DB.DBA.rdf_strlang_impl (in str varchar, in lang any)
{
    lang := cast (lang as varchar);
    if (lang is null)
        signal ('22007', 'Function rdf_strlang_impl needs a valid language ID as its second argument');

    return DB.DBA.RDF_MAKE_LONG_OF_TYPEDSQLVAL (str, null, cast (lang as varchar));
};

We took the initial file from Master branch.

And we made mistake here, because Master Branch contains Virtuoso 6 and not 7.

The 7th version has altogether different DB.DBA.rdf_strlang_impl procedure.
As follows:

create function DB.DBA.rdf_strlang_impl (in str varchar, in lang any)
{
  declare t integer;
  lang := cast (lang as varchar);
  if ((lang is null) or (regexp_match ('^(([a-z][a-z](-[A-Z][A-Z])?)|(x-[A-Za-z0-9]+))\044', lang) is null))
    signal ('22007', 'Function rdf_strlang_impl needs a valid language ID as its second argument');
  if (is_rdf_box (str))
    str := rdf_box_data (str, 1);
  t := __tag (str);
  if (__tag of nvarchar = t)
    str := charset_recode (str, '_WIDE_', 'UTF-8');
  else if (__tag of varchar <> t)
    {
      if (str is null)
        signal ('22007', 'Function rdf_strlang_impl needs a bound value as its first argument, not a NULL');
      str := cast (str as varchar);
    }
  return rdf_box (str, 257, DB.DBA.RDF_TWOBYTE_OF_LANGUAGE (lang), 0, 1);
}
;

So we started to use version 6 procedure with version 7 database.

If this is the cause for our crashes, we are really sorry for spending so much of your time helping us. :(((

Thank you and sorry again,
Stas

P.S.
I will ask further questions on github issue 882.

4

Hi Stas,

You would have seen my colleagues post in github issue 882, thus please continue the thread in the git issue, which given its nature is best pursued there …

5

Hi, Hugh.

Thank you for your help!