I'm currently developing a project in which I use AllegroGraph as graph database. AllegroGraph provides full support for SPARQL 1.1 Query and Update.
I'm trying to write a SPARQL query in which I need some math functions (such as square root and so on). The problem is that every namespace I found isn't working. For example, I tried to use the xpath namespace for the math functions (https://www.w3.org/2005/xpath-functions/math) in this way:
PREFIX func:<http://www.w3.org/2005/xpath-functions/math#>
SELECT ?res
WHERE {
BIND(func:sqrt(4) as ?res) .
}
but it doesn't work since I get the following error:
QUERY FAILED: URI function <http://www.w3.org/2005/xpath-functions/math#sqrt> is not defined
Can you help me to find a workaround please?
Thanks in advance!
Related
I'm trying to write a custom injection for Neovim's Treesitter that will highlight tagged template literals as SQL code.
The code in question looks like this:
import sql from "postgres"
const query = sql` SELECT * FROM my_table`
I'd like to tell Treesitter to parse this tagged template literal as SQL code. This is possible using Treesitter's custom injections; in fact the library uses this same technique to inject syntax for GraphQL tagged template literals, which I believe happens in this file here.
I've written a configuration file at ~/.config/nvim/after/queries/ecma/injections.scm that's quite similar to the GraphQL query, and although Treesitter is recognizing the file, the injection isn't working at all. Does anyone have a solution to this? Ideally, Treesitter would see the sql tagged template literal and interpret the body of the function with the SQL parser.
to make tree sitter use injections in your after/ folder, your injections.scm file should contain ; extends as first line.
Took me many hours to figure it out, it's now documented right here
Similar to what #brotify-force said, but my own SQL injection for Python was not working until I realized what the issue was.
~/.config/nvim/after/queries/python/injections.scm
; extends
(assignment
left: (identifier) #_id (#match? #_id "sql")
right: (string) #sql
)
My original query used #id, but it seems like if you have any captures not starting with an underscore besides the language capture, the Neovim Treesitter machinery gets confused.
I've been trying to do something similar but for template strings that just happen to contain SQL. I was inspired by a VSCode plugin that syntax highlights any template string that has a comment like /*sql*/ in front of it. Here's what I came up with.
; extends
(
(comment) #_comment-text
(#match? #_comment-text "sql")
(template_string) #ql
)
I do wish there was a proper sql module for tree-sitter, and it looks like one is in the works, but the ql module will do for now.
First time using api-platform and Symfony 4 to create an API interface for a MySQL db.
I'm updating an old search interface for the db for which I need to replicate many of the search options. This includes being able to search on a given field using various matching operators/strategies. e.g. starts with, contains exactly equals, etc.
I've set everything up for the api using Annotations.
The #ApiFilter(SearchFilter::class, properties={"fieldname": "strategy"} annotation on my table class works as designed, but I am limited to one-and-only-one strategy per field. I need to be able to pass the strategy to the api search function in the url. something like:
/api/staff?lastname[start]=dav
or
/api/staff?lastname=david&match=contains
or
/api/staff/lastname/son?searchtype=end
would be fine.
I can't figure out how to set this up. Shockingly, to me anyway, this common requirement doesn't seem to be documented at all.
The file CustomSearchFilter.php located at the repo https://github.com/jordonedavidson/custom_search_filter solves this use-case using the
/api/staff?lastname[start]=dav
syntax.
The file was written by Kévin Dunglas (the author of Api Platform) and is presented with his blessing.
I have the following XQuery which I use to fetch documents for a directory.
xquery version "1.0-ml";
cts:search(fn:collection(), cts:directory-query("/Path/To/Docs/", "infinity"))
Now I need to translate this into a REST call but I can't seem to crack it following the documentation on this page.
https://docs.marklogic.com/REST/GET/v1/search
Update:
using the Jersey REST API, It tried this but got 406 Error
String query = "{\"queries\":[ {\"directory-query\":{\"uri\":[\"/Path/to/Docs/\"]},\"infinite\":true} ]}";
String encodedQuery = URLEncoder.encode(query, "UTF-8");
WebTarget target = searchWebTarget.queryParam("structuredQuery", encodedQuery);
final Response response = target.request().get();
Any ideas?
As David said, you don't need to use structured query for this purpose, but in case you have future need:
I believe your original issue was that this is not a well-formed structured query:
{\"queries\":[ {\"directory-query\":{\"uri\":[\"/Path/to/Docs/\"]},\"infinite\":true} ]}
You're missing the top level "query" property. You can find an example of a fully formed structured query that uses directory-query here:
http://docs.marklogic.com/guide/search-dev/structured-query#id_97452
Also, you're probably already aware, but there is a native Java API that sits atop the REST API. You can learn more about this API here:
https://docs.marklogic.com/javadoc/client/index.html
http://docs.marklogic.com/guide/java
Constraining by directory is a query parameter directly on the search API. NO other notation needed.
See the docs here: https://docs.marklogic.com/REST/GET/v1/search
I tried to update embedded triples in marklogic using xquery but it seems to be not working for embedded triples however the same query is working for other triples
can you tell me if there is some other option which needs to specified while performing an update on embedded triples.
The code i used is
xquery version "1.0-ml";
import module namespace sem = "http://marklogic.com/semantics"
at "/Marklogic/semantics.xqy";
let $triples := cts:triples(sem:iri("http://smartlogic.com/document#2012-10-26_DNB.OL_(Citi)_DNB_ASA_(DNB.OL)__Model_Update.61259187.xml"),()())
for $triple in $triples
let $node := sem:database-nodes($triple)
let $replace :=
<sem:triple>
<sem:subject>http://www.example.com/products/1001_Test
</sem:subject>
{$node/sem:predicate, $node/sem:object}
</sem:triple>
return $node ! xdmp:node-replace(., $replace)
My document contains the following triple
<sem:triples xmlns:sem="http://marklogic.com/semantics">
<sem:triple>
<sem:subject>http://smartlogic.com/document#2012-10-26_DNB.OL_(Citi)_DNB_ASA_(DNB.OL)__Model_Update.61259187.xml</sem:subject>
<sem:predicate>http://www.smartlogic.com/schemas/docinfo.rdf#cik</sem:predicate>
<sem:object>datatype="http://www.w3.org/2001/XMLSchema#string</sem:object>
</sem:triple>
</sem:triples>
and i want this particular subject to change into something like this
<sem:subject>http://www.example.com/products/1001_Test</sem:subject>
But when i use the xquery to update it , it does not alter anything, the embedded triple in the documents remains the same.
Because when i tried to see if any of the results have changed to the subject i specified it returned me no results.
I used the following query to test.
SELECT *
WHERE {
<http://www.example.com/products/1001_Test> ?predicate ?object
}
You need to add the option 'all' when you ask for the database nodes backing the triple: sem:database-nodes($triple, 'all').
To be perfectly honest, I am not 100% sure why, but I think this is because your sem:triples element is not the root element of the document it appears on.
I am new to MarkLogic and we are evaluating MarkLogic for our product use case.
We evaluated few NoSQL databases like MongoDB, Couchbase etc.
I am looking for a below type of query search.
(Condition1 OR Condition2) AND (Condition3 OR Condition4) AND (Condition5)
Can MarkLogic provide such type of search query?
I am just started learning MarkLogic and trying to understand the architecture.
Thanks,
Sameer
Yes, MarkLogic provides some high level libraries for this type of functionality. Take a look at Search API.
Start here: https://developer.marklogic.com/learn/2009-07-search-api-walkthrough
And more thorough documentation is here: https://docs.marklogic.com/guide/search-dev/search-api
MarkLogic can handle this kind of logic in many ways as mentioned.
For example, this is how you could setup a search query using the CTS library (I highly recommend the CTS library, since it uses indexes much better, and the construction of them are so much more flexible):
cts:search(//elementName,
cts:and-query((
cts:element-attribute-value-query(xs:QName("entry"), xs:QName("private"), "true"),
cts:or-query((
cts:element-attribute-value-query(xs:QName("entry"), xs:QName("forced"), "false"),
cts:element-attribute-value-query(xs:QName("entry"), xs:QName("forced"), "pending")
))
))
)
This snippet shows both AND and OR logic. The cts:and-query() and cts:or-query() functions can take a list of nodes. The above query says: "Find an element (called elementName) that has an attribute of private='true' AND has either one of the following: forced='true' or forced='pending'".
For much simpler data, you can use xQuery predicates by doing something like the following:
for $node in $xml/parent/child[#param1 eq "test" AND #param2 eq "OK"]/grandchild[#service eq "yahoo" or #service eq "google"]
return $node
The short answer to the original question is "yes". The details of "how" will depend on the approach used to express the queries.
The reference architecture recommends a three-tier approach using the Java or Node.js Client APIs if you use one of those, or HTTP calls to the REST API if you use a different language in your middle tier.
You can also use the Search API (as mentioned by wst) if you're working in MarkLogic's application server (typically as a two-tier architecture). You can do that with either XQuery or Server-side JavaScript, as of MarkLogic 8.