Binding sets of variables when calling sparql update stored query

Support
1

Hi,

I would like to know how I could bind sets of variables when calling a SPARQL UPDATE stored query, instead of binding each variable separatly?

I have searched for a syntax supporting that in both the CLI and the HTTP API documentations. And I haven't find what I am looking for. There is only something about how to bind multiple values for one variable at a time, by separating the values with a blank. But this is not my use case as the values of a set need to be associated together.

In other words, I would like to provide the SPARQL UPDATE the parameters like I would use the VALUES clause as following:

values(?par1 ?par2 ?par3) { ("val1-1" "val2-1" "val3-1") ("val1-2" "val2-2" "val3-2") ... ("val1-n" "val2-n" "val3-n") }

I would have between 1 and 10 sets and not more. It is not a case of bulk load but transactional. But I would like to avoid multiple calls to Stardog from the application to improve the response time.

Thanks
Daniel

2

Daniel,

Neither Stardog nor the SPARQL Update spec facilitate this. As an alternative, you could add the bindings to a temporary named graph, join in your UPDATE and then drop the temporary graph.

Jess

1 Like
3

Thanks Jess for your clear answer.

I will try to group the parameters, that need to be together, in a single variable separated with some special characters (3 or more pipes for example). I will only need to extract each parameter in the stored query. And that should do the job.

Regards
Daniel

4

Can you use the UNNEST function in an update query? It's not standard but might work.

I was also curious if you can use the UNNEST function in an SMS2 mapping. I have a tsv file that has some comma separated fields. (There's a names column with multiple names steve,jill,toby, etc). I was preprocessing it with a python script and then thought, "maybe I can do this with SMS2 if I can use the UNNEST function. Unfortunately that's not the only problem, the file is a mess, so I'll have to stick with python but I was still wondering if you can use that function.

5

All functions can be used with file imports (CSV/JSON). The only problem here is that UNNEST is a SPARQL operator and not a function. :slight_smile: We are aware of the utility here and are keeping in mind the possibility as we evolve our data loading faciltities.

Jess

6

I faced similar limitation of unable to pass list/array of string as a parameter to a stored query.

As a temporary workaround I add a comment as a place holder and replace it in java code with actual

select *

where {

#TO_REPLACE#

?s ?p ?o

}

In java code replace #TO_REPLACE# with

values(?o) { "val1" "val2" “val3" }

I really wish a way to bind list/set/array of things to a query.

7

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.