Hi Holger,
The only reason for using the nestedObjectsList is to get all the
preferred labels of a concept as a single list and in a single json
object instead of getting a different json object for each label.
Basically getting all preferred labels as a list separated by a
delimiter helps to reduce the dependent application's processing time.
Let's say there are 50 preferred labels then after getting the SPARQL
Endpoint service response, splitting those 50 labels by a delimiter
takes less time than calling the get() method 50 times for the 50 json
objects. It is not only about preferred labels. We have alternate
labels, ambiguous alternate labels, etc for our concepts.
When using the nestedObjectsList to get all the labels of a
concept as list then since the delimiter is comma for the list of
labels so when a label itself contains comma then splitting the labels
by the delimiter comma fails to give the correct labels.
_SPARQL response *with* nestedObjectsList in the SPARQL*
*_Below SPARQL Endpoint service response contains single json object
for all the preferred labels of the concept Africa._*
*_
_SPARQL response *without* nestedObjectsList in the SPARQL*
*_
Below SPARQL Endpoint service response contains different json objects
for the preferred labels of the concept Arfica.
_*_*
*_*_
Thanks,
Sanjeev
On Wednesday, February 21, 2018 at 7:18:45 AM UTC+5:30, Holger
Knublauch wrote:
Hi Sanjeev,
On 20/02/2018 20:42, sanjeev devireddy wrote:
Hi Holger,
Your suggestion of using built-in str function works only
when all the language labels are queried. As per our requirement,
we wrote a SPARQL (please check below) to get English language
labels only and in this case the built-in str function
fails(please check below screenshot) to convert a language-tagged
string literal to a simple xsd:string literal. Could you please
check the below SPARQL to see if there is any change that can be
done to the SPARQL to convert a language-tagged string literal to
a simple xsd:string literal?
_
SPARQL:
SELECT DISTINCT ?preferredlabel ?stringPrefLabel ?result
WHERE {
GRAPH <urn:x-evn-master:geo> {
{
?result a
<http://topquadrant.com/ns/examples/geography#Continent>
<http://topquadrant.com/ns/examples/geography#Continent> .
} .
BIND (search:nestedObjectsList(?result, skos:prefLabel,
"result", ?none, "en") AS ?preferredlabel) .
BIND (str(?preferredlabel) AS ?stringPrefLabel) .
}
}
ORDER BY (LCASE(?label))
_
in your example data, the nestedObjectsList function produces
strings such as "Asia (en)" that use a non-standard format of
saving language tags. In contrast, the official way looks (in
Turtle) like "Asia"@en - the string literal itself does not
include the language but the tag is a special attachment to the
literal. In the former case, the str function already operates on
an xsd:string without language tag - it doesn't use the ... (en)
naming convention. If you want to get rid of the " (en)
sub-strings, you could for example use a REPLACE, e.g.
BIND (REPLACE(?preferredlabel, " \\(en\\)", "") AS ?s)
Coming to the other question on dealing with the label that has
comma in it,it seems in the above post I was referring to a bad
example so below is another example that I wan t to share.
Our taxonomy has a concept and it's preferred labels in English &
French languages are *Government, Central/Federal (en)* and
*Gouvernement, Central/Fédéral (fr)*. Here the thing to observe
is that each label contains comma in it. Now, when a SPARQL is
written to get the above preferred labels of the concept and use
SPARQL Endpoint service then the response contains the two
language labels separated by comma as shown below. Now the
challenge is that the labels contains a comma in them and the
separator for the English and French labels is also a comma.So,
using comma as separator to get labels from the below json will
give the 4 values i) Government*, *ii)Central/Federal (en)*,*
iii) Gouvernement , iv)Central/Fédéral (fr). But the requirement
is to get the actual English and French language labels correctly.
e.g.: "prefLabel_0":{"value":"*Government, Central/Federal (en)*,
*Gouvernement, Central/Fédéral (fr)*","type":"literal"}
So, we just want to check that what could be the best way in this
kind of scenarios to get the labels correctly? Is there way to
specify a separator(like pipe (|) ) for the preferred/ alternate
labels list in the above json through the SPARQL Endpoint service
or so?
Why do you need to go through nestedObjectsList at all? This would
give you only the french labels:
SELECT *
WHERE {
?result a g:Continent .
?result skos:prefLabel ?label .
FILTER (lang(?label) = "fr")
}
Holger
Thanks,
Sanjeev
On Tuesday, February 20, 2018 at 5:11:53 AM UTC+5:30, Holger
Knublauch wrote:
On 19/02/2018 23:59, sanjeev devireddy wrote:
Hi,
We have the following two challenges while
fetching/parsing the list of preferred/alternative labels
received from the the SPARQL Endpoint service response.
1)To get the preferred/alternative label(s) without language
tags. We just want to check that is there a way to get the
labels without language tags?
You can convert a language-tagged string literal to a simple
xsd:string literal using the built-in str function, e.g.
SELECT *
WHERE {
?concept skos:prefLabel|skos:altLabel ?label .
BIND (str(?label) AS ?stringLabel) .
}
2)When there is a comma in the labels of
preferred/alternative labels (please check the below example).
Assume that there a single concept named Latin America,
North America. As shown below, we can see that there are two
labels of the languages English & French. Here we can
observe that in each language label there is a comma. In
this case, using comma as a separator will fail to get the
English & French language labels of a single concept Latin
America, North America.
"prefLabel_0": { "type": "literal" , "value": "Latin
America, North America(en), Amérique latine, Amérique du
Nord(fr)" }
So we just want to check what could be the best way in
this kind of scenario?
So you want the query to return all sub-strings as separated
by commas? The following example will produce an iteration
over all substrings, and trim them off extra spaces in case
the string is irregular.
SELECT *
WHERE {
BIND ("Latin America, North America(en), Amérique latine,
Amérique du Nord(fr)" AS ?str)
?sub spif:split (?str ",") .
BIND (spif:trim(?sub) AS ?trimmed) .
}
To get rid of the language tags, you'd need to do further
string processing using functions like spif:indexOf and SUBSTR.
HTH
Holger
--
You received this message because you are subscribed to the
Google Groups "TopBraid Suite Users" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to [email protected] <javascript:>.
For more options, visit https://groups.google.com/d/optout
<https://groups.google.com/d/optout>.
--
You received this message because you are subscribed to the Google
Groups "TopBraid Suite Users" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to [email protected]
<mailto:[email protected]>.
For more options, visit https://groups.google.com/d/optout.
