February 2026 - BaseX-Talk - mailman.uni-konstanz.de

job:execute and locks
by Graydon Saunders 27 Feb '26

27 Feb '26

Hello -- I'm on the 12.3 413f355 beta and using the BaseX GUI. So this query runs forever in a way that makes me think I've got a locking problem: declare namespace v2="https://schema.iccsafe.org/book/schema/1.0"; (: list of query names, in order; order is important here :) declare variable $queryNameList as xs:string+ := ('load-before-1.xq', 'load-after-2.xq','wordcount-3.xq'); (: job:execute wants URIs. We could use file:resolve-path here but minimalist relative URIs work :) declare variable $queryURIList as xs:anyURI+ := $queryNameList ! xs:anyURI(.); (: the map with the file paths :) declare variable $namesMap as xs:string := '/home/graydon/work/tests/compare_map.xquery'; (: where reports should go :) declare variable $reportsDir as xs:string := '/home/graydon/work/tests/reports/'; (: try putting path pairs in an external map so we ought not have any db lock issues :) let $names as map(xs:string,map(xs:string,xs:string)) := file:read-text($namesMap) => xquery:eval() return (: bindings to pass values to the queries we're about to run :) for key $k value $v in $names let $bindings as map(*) := {'reportsDir': $reportsDir, 'sourceName': $v?source, 'resultName': $v?result } return for $query in $queryURIList return job:execute($query,$bindings) When I could get the information to set the values in $v?source and $v?result without having to use doc() to look inside files, this query's immediate ancestor worked. The docs at https://docs.basex.org/main/Job_Functions#job:execute say "It is safe to use the function if the calling query does not lock anything." Does having opened a document via doc() in another query in the GUI leave that document locked? If so, is there a way to drop those locks? Am I locking something with this query? (The "generate a bxs script" version (from the same map!) runs in about 45 seconds, so I think the queries being called via job:execute work.) Thanks! Graydon

2 6

A modifying Identity Transform pattern for JSON via XQuery 4.0
by ykhabins＠bellsouth.net 27 Feb '26

27 Feb '26

Hello, A modifying XSLT Identity Transform pattern is a powerful and elegant technique that uses a general "catch-all" template to copy the entire input XML structure, combined with more specific templates that override the default behavior to make targeted changes. This approach ensures that everything not explicitly matched by a custom template remains unchanged. Is it possible to do the same in BaseX 12 for JSON via XQuery 4.0? Is it possible to implement the XQuery 4.0 & JSON equivalent of XSLT’s identity template, but with the ability to make small, targeted changes while otherwise copying the input unchanged: 1) Parse JSON to XQuery’s native JSON model (order‑preserving maps/arrays) using the JSON module’s w3 format (lossless, W3C‑aligned). 2) Walk that value and rebuild it, applying tweaks where needed (rename a key, drop a field, change a value, etc.), while copying everything else. 3) Serialize back to JSON with the same format (and any serialization parameters that are required). Regards, Yitzhak Khabinsky

2 1

Very suprising results by combining BaseX, Saxon, XSLT-Packages and a custom Saxon-Initializer
by Nico Kutscherauer 26 Feb '26

26 Feb '26

Hi BaseX team, I admit, I'm pushing the combinations a bit, but otherwise it wouldn't be any fun... ;-) I want to call an XSLT stylesheet with Saxon in BaseX. However, the stylesheet should use an XSLT package. For XSLT packages, I wrote a Saxon initialiser some time ago that automatically recognises packages in the classpath and configures Saxon so that they are available. When embedding it in BaseX, however, I now encounter very strange behaviour when I execute the stylesheet twice (or more) in the same session. I have reproduced this in this GitHub project: https://github.com/nkutsche/basex-saxon-xslt-package-issue When you run either the shell script basex or basex.bat in the project, the following happens: 1. The Maven project fetches the necessary dependencies. - Saxon 12.9 & BaseX 12.2 - xpath-model - the XSLT package I want to use in my XSLTs (https://github.com/nkutsche/xpath-model/) - xslt-pkg-manager120 - the Saxon initialiser mentioned above 2. The system property SAXON_INITIALIZER is set to the Saxon initialiser 3. The XQuery script xquer.xq is executed with BaseX: - this executes test.xsl twice on test.xml with xslt:transform() - the sequence of both transformation results is the result of the main query. 4. test.xsl uses a function from the XPath model: the XPath expression in test.xml is parsed into an XML tree. The second result is then very surprising: ---- result 1 ---- <expr><operation type="step"><slash/><arg><locationStep axis="child"><nodeTest kind="node"/><predicate><integer value="1"/></predicate></locationStep></arg><slash/><arg><locationStep axis="child"><nodeTest kind="node"/><predicate><integer value="4"/></predicate></locationStep></arg></operation></expr> ---- result 2 ---- <expr><XPath><Expr Predicate="sequence"><EOF><ExprSingle><Expr Predicate="or"><EOF><Expr Predicate="and"><EOF><ComparisonExpr><Expr Predicate="concat"><EOF><Expr Predicate="range"><EOF><Expr Predicate="additive"><EOF><Expr Predicate="multiplicativ"><EOF><Expr Predicate="union"><EOF><Expr Predicate="intersect-except"><EOF><Expr Predicate="instance-of"><EOF><Expr Predicate="treat-as"><EOF><Expr Predicate="castable"><EOF><Expr Predicate="cast"><EOF><Expr Predicate="arrow"><EOF><Expr Predicate="unary"><EOF><ValueExpr><Expr Predicate="map"><EOF><Expr Predicate="step"><EOF><TOKEN>/</TOKEN></EOF><EOF><Expr Predicate="step"><EOF>[...] The complete output can also be found here: https://github.com/nkutsche/basex-saxon-xslt-package-issue/actions/runs/224… The first result is as expected. I can hardly explain the second one. Except that the elements look similar to my intermediate results, but as far as I know, they never occur in this combination. My only guess is that a few templates in the package suddenly are no longer matching or no longer match as they should. Do you have any idea where this might be coming from? I couldn't reproduce the behaviour when I called Saxon directly twice in a JVM. The problem could still be in Saxon, of course, but I would first have to reproduce it without BaseX. Thank you for your help! Best Regards, Nico -- Nico Kutscherauer XML Developer data2type GmbH Im Breitspiel 6A 69126 Heidelberg Tel.: +49-(0)6221-73 912 60 www.data2type.de <https://www.data2type.de> Workingdays: Tue/Wed/Thu/Fri(HO) Workinghours: 10-18.30 Company’s head office: Heidelberg Commercial Register court: AG Mannheim HRB 715195 CEO: Manuel Montero Pineda

2 2

test
by ykhabins＠bellsouth.net 23 Feb '26

23 Feb '26

Test Regards, Yitzhak Khabinsky

1 0

db:node-id() and updating operations
by Graydon Saunders 18 Feb '26

18 Feb '26

Hello -- Using db:node-id() is an order of magnitude faster than using generate-id() to relate various nodes to their ancestors when constructing a composite id. This works great. Now I have a map (the keys are the value returned by db:node-id() for specific elements) and I want to update either a document or the db to have the new, calculated values as id attribute values. I expected that I could use for $thisDoc in db:get('ID-DB') ...calculate some stuff... ... $hasIDKey is the map:keys() of the map of calculated values... return $thisDoc update {(//*[db:node-id(.) = $hasIDKey]) ! (insert node attribute { 'id' } { $composedMap(./db:node-id(.))?composed } into .)} This pattern worked with generated-id() values; with db:node-id() the error I get is: [db:node] No database node: <publication.../>. (Publication is the expected document element.) db:node-id() worked in a function in a library module when $thisDoc was a parameter, so I'm at a bit of a loss; I would have expected that db:node-id() is the stable identifier and could work with updating expressions but perhaps that was optimism. Is there an appropriate way to relate a db:node-id() value to an element in an updating expression? Thanks! Graydon

2 3

Max message size in Websockets exceeded?
by Ben Engbers 18 Feb '26

18 Feb '26

Hi, According to "https://docs.basex.org/11/WebSockets", the default limit for messages is 64 Kb. Exceeding the limit should raise an exception. However when using my BasexCpp library on a fast machine with a lot of memory, I can add successfuly a file, sized 358 kB, with one add-message. Adding the same file to a basexserver on a remote, slow machine, results in a crash. Neither the local nor remote console report a message. I downloaded the Basex-sources and found in org.basex.core.cmd/Add.java that the cache() function may use a ADDCACHE option. Since the memory size of the server is limited, I wondered if setting this option might force the server to explicitly use a disk-cache. Does the Basex server protocol offers any features for setting this option? I have been collecting information on websocket programming and learned a lot on handshaking, maaual paring, framing, event-loops and other concepts. My guess is that most is not very helpfull. Does anyone have good sugestions for relevant information? Ben -- Ben Engbers Grietjeshof 77 6721 VH Bennekom +31 6 23634840

3 4

persistent id values?
by Graydon Saunders 16 Feb '26

16 Feb '26

Hi! I'm loading a document with doc() and doing a bunch of stuff to regenerate id attribute values; I'm using generated ids (from generate-id()) as map keys so I can generate the local, this-element id component and then put them all together from the list of ancestor ids. That works. What doesn't work is trying to use those ids to identify the element which should get this newly constructed id value; calling generate-id() again on the same document doesn't give the same answers. (which is to be expected, at least as I understand how generate-id() is defined.) I have this vague memory that there's a BaseX function that will produce a stable id, but can't remember what it is called. (Or if it will work when I get the document with doc().) Is there such a thing, and if so, what module should I be looking in? Thanks! Graydon

4 4

Re: function as default parameter
by Gunther Rademacher 14 Feb '26

14 Feb '26

1 0

function as default parameter
by Rob Stapper 14 Feb '26

14 Feb '26

Good morning, I came across some unexpected behaviour when using a function as a default functionparameter: declare function local:test ( $exec as function(*)? := fn( $a, $b) { $a => trace("test1: ") } ) { $exec( "a", "b") } ; local:test(()) I would expect a trace- and a execution result, but I got nothing. Are my expectations wrong or could it be a bug? PS code snipped added as test.xqm PS The optimized query shows an empty value. mvgr. Rob Stapper

1 0

aligning sequences of text?
by Graydon Saunders 12 Feb '26

12 Feb '26

Hello! If I have two (fairly long) sequences of text, ('The', 'words', 'are', 'sequence', 'members') and I want all the index numbers of matching pairs despite the sequences only mostly matching (so a word, or several words, can be missing from sequence A or sequence B), is there an established algorithm for doing this? (If I search on "aligning sequences" I get bioinformatics about gene sequences; if I search on "aligning text" I get typography.) Thanks! Graydon

4 5