Hi all - I think that I weak grasp of the practical requirements, but I'm unable to find a reference in the spec to clarify what's happening in the following:
I have a constructor: return( for $item in $file-list let $item-less := replace($item, '^\d{1,}-', '') (: returns "missing" supplemental files :) where ($item-less[(not(. = ($files-titles, $excludes)))]) or ($item-less[(. = $files-titles)]) return( <relatedItem> <title>{$item-less}</title> <mimeType> {if ($item-less = $files-titles) then ($content/files/file/name[. = $item-less]/following-sibling::mimetype/text()) else (fetch:content-type(concat(file:parent(document-uri($doc)),$item)))} </mimeType> <uri>{file:parent(document-uri($doc))}</uri> </relatedItem> ) )
and it works fine so long at the <relatedItem> is not namespaced. As soon as the <relatedItem> is namespaced; e.g. <relatedItem xmlns=" http://bridger.net">, then I need to use the following XPath in the `then` keyword:
($content/*:files/*:file/*:name[. = $item-less]/following-sibling::*:mimetype/text())
The input documents (in the SSCCE, under sample-data/*/metadata.xml) are not in a namespace.
Would someone mind providing some context for the "why" here? I've been looking through Priscilla Walmsley's 2nd edition, and the specification, but haven't been able to find an answer. I'm almost certainly looking for the wrong thing(s).
Any help will be greatly appreciated! Thank you in advance. Best, Bridger
Why not attach that SSCCE, too? :)
Apologies, Bridger
On Thu, Dec 29, 2016 at 6:29 PM, Bridger Dyson-Smith bdysonsmith@gmail.com wrote:
Hi all - I think that I weak grasp of the practical requirements, but I'm unable to find a reference in the spec to clarify what's happening in the following:
I have a constructor: return( for $item in $file-list let $item-less := replace($item, '^\d{1,}-', '') (: returns "missing" supplemental files :) where ($item-less[(not(. = ($files-titles, $excludes)))]) or ($item-less[(. = $files-titles)]) return( <relatedItem> <title>{$item-less}</title> <mimeType> {if ($item-less = $files-titles) then ($content/files/file/name[. = $item-less]/following-sibling: :mimetype/text()) else (fetch:content-type(concat(file:parent(document-uri($doc) ),$item)))} </mimeType> <uri>{file:parent(document-uri($doc))}</uri> </relatedItem> ) )
and it works fine so long at the <relatedItem> is not namespaced. As soon as the <relatedItem> is namespaced; e.g. <relatedItem xmlns=" http://bridger.net">, then I need to use the following XPath in the `then` keyword:
($content/*:files/*:file/*:name[. = $item-less]/following-sibling: :*:mimetype/text())
The input documents (in the SSCCE, under sample-data/*/metadata.xml) are not in a namespace.
Would someone mind providing some context for the "why" here? I've been looking through Priscilla Walmsley's 2nd edition, and the specification, but haven't been able to find an answer. I'm almost certainly looking for the wrong thing(s).
Any help will be greatly appreciated! Thank you in advance. Best, Bridger
Hi Bridger,
Sorry for letting you wait:
and it works fine so long at the <relatedItem> is not namespaced.
I tried to further strip down your code. Do you think that the following examples are equivalent?
WORKS:
let $content := document { <root>BLA</root> } return <item>{ $content/root }</item>
DOES NOT WORK:
let $content := document { <root>BLA</root> } return <item xmlns='ns'>{ $content/root }</item>
This is indeed compliant with the spec. The hint may be a bit hidden, but the spec states that “[…] the default element/type namespace may be affected by namespace declaration attributes found inside the element constructor.”.
I agree that the decision is somewhat counterintuitive, because you would expect that moving your code to a different place will not change its behavior.
Hope this helps, Christian
[1] https://www.w3.org/TR/xquery-31/#id-element-constructor
The input documents (in the SSCCE, under sample-data/*/metadata.xml) are not in a namespace.
Would someone mind providing some context for the "why" here? I've been looking through Priscilla Walmsley's 2nd edition, and the specification, but haven't been able to find an answer. I'm almost certainly looking for the wrong thing(s).
Any help will be greatly appreciated! Thank you in advance. Best, Bridger
Hi Christian,
On Mon, Jan 2, 2017 at 5:55 AM, Christian Grün christian.gruen@gmail.com wrote:
Hi Bridger,
Sorry for letting you wait:
It's no trouble at all -- thank you for your time and trouble.
and it works fine so long at the <relatedItem> is not namespaced.
I tried to further strip down your code. Do you think that the following examples are equivalent?
WORKS:
let $content := document { <root>BLA</root> } return <item>{ $content/root }</item>
DOES NOT WORK:
let $content := document { <root>BLA</root> } return <item xmlns='ns'>{ $content/root }</item>
HA! That's a *much* better expression of the problem - thank you for
condensing it!
This is indeed compliant with the spec. The hint may be a bit hidden, but the spec states that “[…] the default element/type namespace may be affected by namespace declaration attributes found inside the element constructor.”.
I agree that the decision is somewhat counterintuitive, because you would expect that moving your code to a different place will not change its behavior.
Hope this helps, Christian
Absolutely. Again, thank you kindly for your help!
Happy New Year. Best, Bridger
[1] https://www.w3.org/TR/xquery-31/#id-element-constructor
The input documents (in the SSCCE, under sample-data/*/metadata.xml) are
not
in a namespace.
Would someone mind providing some context for the "why" here? I've been looking through Priscilla Walmsley's 2nd edition, and the specification,
but
haven't been able to find an answer. I'm almost certainly looking for the wrong thing(s).
Any help will be greatly appreciated! Thank you in advance. Best, Bridger
basex-talk@mailman.uni-konstanz.de
-
Bridger Dyson-Smith -
Christian Grün