order clause is not produce an ordered output

[andrewufrank]

i use the query
SELECT ?o (count( ?s) as ?count) WHERE { ?s http://auf.us.to/nlp2015/lemmaPoS ?o } group by ?o ORDER by desc (?count) LIMIT 100

EOQ

in 4s-query (same result in 4s-http) and it produces the correct list, but not ordered:
4store>SELECT ?o (count( ?s) as ?count) WHERE { ?s http://auf.us.to/nlp2015/lemmaPoS ?o } group by ?o ORDER by desc (?count) LIMIT 100

EOQ

?o ?count
"worthy:JJ"@us "2"^^http://www.w3.org/2001/XMLSchema#integer
"store-room:NN"@us "1"^^http://www.w3.org/2001/XMLSchema#integer
"thyself:RB"@us "1"^^http://www.w3.org/2001/XMLSchema#integer
"throughout:IN"@us "4"^^http://www.w3.org/2001/XMLSchema#integer
"wound:VB"@us "9"^^http://www.w3.org/2001/XMLSchema#integer
"cheese:NN"@us "17"^^http://www.w3.org/2001/XMLSchema#integer
"thicker:JJ"@us "1"^^http://www.w3.org/2001/XMLSchema#integer
"leisurely:JJ"@us "1"^^http://www.w3.org/2001/XMLSchema#integer
...

what is wrong wtih the query?
(ps. how do i avoid the clutter and have the result as "1" only?)

thank you
andrew

[swh]

Good question... looks like the query engine isn't treating the numbers as real xsd:integers - sort order for unknown datatypes is undefined, and if it thought they were xsd:integers they would be rendered as bare-word integers.

Can you show a sample of the input data?

[andrewufrank]

here is a small example (not terribly small).
it imports with 4-import ...

thank you for your efforts!
andrew

Prof. Dr. sc.techn. Dr. h.c. Andrew U. Frank
+43 1 58801 12710 direct
Geoinformation, TU Wien +43 1 58801 12700 office
Gusshausstr. 27-29 +43 1 55801 12799 fax
1040 Wien Austria +43 676 419 25 72 mobil

On 09/28/2015 11:53 AM, Steve Harris wrote:

Good question... looks like the query engine isn't treating the numbers
as real xsd:integers - sort order for unknown datatypes is undefined,
and if it thought they were xsd:integers they would be rendered as
bare-word integers.

Can you show a sample of the input data?

—
Reply to this email directly or view it on GitHub
#133 (comment).

[swh]

Is the example attached to the ticket? I can't see it.

4s-import will import anything that's legal RDF, and you can write things like "VII"^^my:romanNumerals, though it won't sort properly, and you can't do arithmetic with it.

[andrewufrank]

thank you for your hint. i prepared a reduced file for attachement and found that the error is gone - it sorts and the number are shown as 1, 34 etc.
any hints for what i should search to find the offending data?
the value for count is not from my data but produced internally, why can there be a confusion (even if my data is wrong. it counted the number of triples apparently correct)?
thank you !

[swh]

I don't know I'm afraid... I would guess that the datatype URI is subtly wrong in the source data - but it looked OK to me.

You could do a search that returned the datatype URI of the objects of that predicate?

[swh]

Ah! Sorry, I've just seen that it's the output of the COUNT() aggregate that not being sorted correctly.

This might be a known bug (I remember there was something in this area), but it should be fixable.

[andrewufrank]

i was however surprised, that the bug disappeared with less data. i will
check the data, but it does not contain any datatype specification other
than @us in the turtle input.
thank you for your attention to this problem!
andrew

Prof. Dr. sc.techn. Dr. h.c. Andrew U. Frank
+43 1 58801 12710 direct
Geoinformation, TU Wien +43 1 58801 12700 office
Gusshausstr. 27-29 +43 1 55801 12799 fax
1040 Wien Austria +43 676 419 25 72 mobil

On 09/28/2015 04:49 PM, Steve Harris wrote:

Ah! Sorry, I've just seen that it's the output of the COUNT() aggregate
that not being sorted correctly.

This might be a known bug (I remember there was something in this area),
but it should be fixable.

—
Reply to this email directly or view it on GitHub
#133 (comment).

[swh]

Ah, interesting.

So in that case it's definitely a bug in the engine, not the data.

I can't offhand think of any reason why it might behave differently with more/less data.