|
Subject: Re: [xsl] How to write (existential) predicates with maps/Why is there no effective boolean value for a map? From: "Michael Kay mike@xxxxxxxxxxxx" <xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> Date: Tue, 12 Feb 2019 16:41:51 -0000 |
I've improved the error message so it now says:
FORG0006: Effective boolean value is not defined for sequence starting with a
map
(map{"name":"foo", })
The reason for the problem:
> parse-json($json)?locations?*[?types?*[?name = 'foo']]"
locations?* => a sequence of maps
?types => a sequence of arrays
?types?* => a sequence of maps
So the value of the outer predicate is a sequence of zero-or-more maps, and
you can't get the EBV of a map.
Created W3C test case predicate-056.
>What is the reason that the effective boolean value was not extended to give
true for a sequence with a map?
Sentiment in the WGs slowly moved away from weak typing and implicit
conversion over the years that followed XPath 1.0 (Getting arrays to be
atomizable was a bit of a battle). I think the richer the type system becomes,
the more useful it is to have errors rather than implicit conversions. I saw a
horrible one last week in which someone did <xsl:with-param>2</xsl:with-param>
and then tried to use the value as a "numeric" predicate, not realising it was
actually a node.
When I started this I was fairly neutral about it, but I've now used
Javascript enough to form a strong distaste for weak typing.
Michael Kay
Saxonica
> On 12 Feb 2019, at 16:07, Martin Honnen martin.honnen@xxxxxx
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>
> On 12.02.2019 15:40, Michael Kay mike@xxxxxxxxxxxx wrote:
>> I'm having trouble understanding/reproducing this. Can you supply a
complete repro? (I.e., the source data that results in this error)
>
>
> The example JSON is e.g.
>
> {
> "locations" : [
> {
> "id" : "i1",
> "types" : [
> {
> "name" : "foo"
> },
> {
> "name" : "bar"
> }
> ]
> },
> {
> "id" : "i2",
> "types" : [
> {
> "name" : "baz"
> }
> ]
> },
> {
> "id" : "i3",
> "types" : [
> {
> "name" : "foo"
> },
> {
> "name" : "baz"
> }
> ]
> }
> ]
> }
>
> If you pass it to the parse-json function and use the result as the context
item to the three XPath expressions I have posted then the first one gives
that error while the other two return two maps.
>
> A complete repro in XSLT is
>
> <?xml version="1.0" encoding="UTF-8"?>
> <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
> xmlns:xs="http://www.w3.org/2001/XMLSchema"
> exclude-result-prefixes="#all"
> version="3.0">
>
> <xsl:output method="adaptive"/>
>
> <xsl:param name="json" as="xs:string">
> {
> "locations" : [
> {
> "id" : "i1",
> "types" : [
> {
> "name" : "foo"
> },
> {
> "name" : "bar"
> }
> ]
> },
> {
> "id" : "i2",
> "types" : [
> {
> "name" : "baz"
> }
> ]
> },
> {
> "id" : "i3",
> "types" : [
> {
> "name" : "foo"
> },
> {
> "name" : "baz"
> }
> ]
> }
> ]
> }
> </xsl:param>
>
> <xsl:template match="/" name="xsl:initial-template">
> <xsl:sequence select="parse-json($json)?locations?*[?types?*[?name =
'foo']]"/>
> </xsl:template>
>
> </xsl:stylesheet>
>
>
>
>
>>> On 11 Feb 2019, at 12:09, Martin Honnen martin.honnen@xxxxxx
<xsl-list-service@xxxxxxxxxxxxxxxxxxxxxx> wrote:
>>>
>>> When using XPath 3.1 (e.g. in XSLT 3) on maps I have found that I have to
change my coding habit a bit when writing predicates that want to check the
existence of some nested map, while I hoped to be able to write e.g.
>>>
>>> ?locations?*[?types?*[?name = 'foo']]
>>>
>>> to select all (map) members of the "locations" array that have a "types"
array with at least one (map) member having a property "name" with value "foo"
I get an error
>>>
>>> Effective boolean value is not defined for sequence starting with an
atomic value other than a boolean, number, or string
>>>
>>>
>>> So in contrast to my experience with writing predicates on XML it seems
for maps I have to explicitly use the "exists" function e.g.
>>>
>>> ?locations?*[exists(?types?*[?name = 'foo'])]
>>>
>>> or a "some .. in" expression
>>>
>>> ?locations?*[some $m in ?types?* satisfies $m?name = 'foo']
>>>
>>>
>>> Is there any more compact way to write such a check?
>>>
>>> What is the reason that the effective boolean value was not extended to
give true for a sequence with a map?
| Current Thread |
|---|
|
| <- Previous | Index | Next -> |
|---|---|---|
| Re: [xsl] How to write (existential, Martin Honnen martin | Thread | Re: [xsl] How to write (existential, Liam R. E. Quin liam |
| Re: [xsl] How to write (existential, Martin Honnen martin | Date | Re: [xsl] How to write (existential, Liam R. E. Quin liam |
| Month |