Re: [basex-talk] losses von tags?
Am 01.11.2017 um 22:54 schrieb michael4...@arcor.de:
Hello,
There is a *problem in transforming data vom XML in HTML*.
In XML there are HTML-tags included.
XML original: *text is important.*
HTML goal: *text is important.*
(1) result: text is important.
let $data := text is important.
return
{$data/*data()*}
(2) result: text is important.
let $data := text is important.
return
{$data/*text()*}
(3) result: text is important.
let $data := text is important.
return
{*$data*}
Is there another possibility to get the result desired?
Use $data/node() to select all child nodes of the "dat" element.
[basex-talk] losses von tags?
Hello,
There is a problem in transforming data vom XML in HTML.
In XML there are HTML-tags included.
XML original: text is important.
HTML goal: text is important.
(1) result: text is important.
let $data := text is important.
return
{$data/data()}
(2) result: text is important.
let $data := text is important.
return
{$data/text()}
(3) result: text is important.
let $data := text is important.
return
{$data}
Is there another possibility to get the result desired?
Thanking you,
Michael
Re: [basex-talk] Rounding/parsing decimal vs float
Awesome Leo, thanks!
Similar to what I was thinking about but with a lot or after-thought
optimization!
Great suggestion.
M.
Il 01 nov 2017 1:59 PM, "Leonard Wörteler" <
leonard.woerte...@uni-konstanz.de> ha scritto:
> Hi Marco,
>
> I also do not know of a built-in way to do this, but here is my shot at an
> implementation in XQuery. It uses fast exponentiation [1] and throws a
> custom error if there is more than one "e" in the input.
>
> Hope that helps (or is of interest), I had fun hacking it together,
> Leo
>
> declare function local:pow($n, $k) {
> if($k eq 0) then xs:decimal('1')
> else if($k eq 1) then $n
> else (
> let $d := $k idiv 2,
> $m := $k mod 2
> return if($m eq 0) then local:pow($n * $n, $d)
> else local:pow($n * $n, $d) * $n
> )
> };
>
> declare function local:parse-decimal($str as xs:string) as xs:decimal {
> let $parts := tokenize($str, '[eE]')
> return switch(count($parts))
> case 1 return xs:decimal($str)
> case 2 return (
> let $exp := xs:long($parts[2]),
> $base := xs:decimal($parts[1]),
> $b10 := if($exp gt 0) then xs:decimal('10') else
> xs:decimal('0.1')
> return $base * local:pow($b10, $exp)
> )
> default return error(xs:QName('local:NODECIML'), 'Wrong format: '
> || $str)
> };
>
> local:parse-decimal('123.45e-300')
>
> [1] https://en.wikipedia.org/wiki/Exponentiation_by_squaring
>
> Am Mittwoch, 01. November 2017 13:21 CET, Marco Lettere <
> m.lett...@gmail.com> schrieb:
> > Thanks Kristian,
> > unfortunately this introduces rounding errors which I'm trying to avoid
> > because managing units, measures and monetary amounts.
> > M.
> >
> > Il 01 nov 2017 12:57 PM, "Kristian Kankainen" ha
> > scritto:
> >
> > You need to use xs:float or xs:double instead of xs:decimal to be able to
> > use the 'e' or 'E' as the exponent separator.
> >
> > Br,
> > Kristian K1. nov 2017 13:18 kirjutas kuupäeval Marco Lettere <
> > m.lett...@gmail.com>:
> > >
> > > Hi all,
> > >
> > > I thought of asking this in parallel of hacking my own parsing
> procedure
> > > ...
> > >
> > > Is there a native way to parse scientific notation string into
> > > xs:decimal since a direct casting (xs:decimal("1e1")) is not allowed?
> > >
> > > Thanks,
> > >
> > > Marco.
> > >
>
>
Re: [basex-talk] ?==?utf-8?q? Rounding/parsing decimal vs float
Hi Marco,
I also do not know of a built-in way to do this, but here is my shot at an
implementation in XQuery. It uses fast exponentiation [1] and throws a custom
error if there is more than one "e" in the input.
Hope that helps (or is of interest), I had fun hacking it together,
Leo
declare function local:pow($n, $k) {
if($k eq 0) then xs:decimal('1')
else if($k eq 1) then $n
else (
let $d := $k idiv 2,
$m := $k mod 2
return if($m eq 0) then local:pow($n * $n, $d)
else local:pow($n * $n, $d) * $n
)
};
declare function local:parse-decimal($str as xs:string) as xs:decimal {
let $parts := tokenize($str, '[eE]')
return switch(count($parts))
case 1 return xs:decimal($str)
case 2 return (
let $exp := xs:long($parts[2]),
$base := xs:decimal($parts[1]),
$b10 := if($exp gt 0) then xs:decimal('10') else
xs:decimal('0.1')
return $base * local:pow($b10, $exp)
)
default return error(xs:QName('local:NODECIML'), 'Wrong format: ' ||
$str)
};
local:parse-decimal('123.45e-300')
[1] https://en.wikipedia.org/wiki/Exponentiation_by_squaring
Am Mittwoch, 01. November 2017 13:21 CET, Marco Lettere
schrieb:
> Thanks Kristian,
> unfortunately this introduces rounding errors which I'm trying to avoid
> because managing units, measures and monetary amounts.
> M.
>
> Il 01 nov 2017 12:57 PM, "Kristian Kankainen" ha
> scritto:
>
> You need to use xs:float or xs:double instead of xs:decimal to be able to
> use the 'e' or 'E' as the exponent separator.
>
> Br,
> Kristian K1. nov 2017 13:18 kirjutas kuupäeval Marco Lettere <
> m.lett...@gmail.com>:
> >
> > Hi all,
> >
> > I thought of asking this in parallel of hacking my own parsing procedure
> > ...
> >
> > Is there a native way to parse scientific notation string into
> > xs:decimal since a direct casting (xs:decimal("1e1")) is not allowed?
> >
> > Thanks,
> >
> > Marco.
> >
Re: [basex-talk] Rounding/parsing decimal vs float
Thanks Kristian,
unfortunately this introduces rounding errors which I'm trying to avoid
because managing units, measures and monetary amounts.
M.
Il 01 nov 2017 12:57 PM, "Kristian Kankainen" ha
scritto:
You need to use xs:float or xs:double instead of xs:decimal to be able to
use the 'e' or 'E' as the exponent separator.
Br,
Kristian K1. nov 2017 13:18 kirjutas kuupäeval Marco Lettere <
m.lett...@gmail.com>:
>
> Hi all,
>
> I thought of asking this in parallel of hacking my own parsing procedure
> ...
>
> Is there a native way to parse scientific notation string into
> xs:decimal since a direct casting (xs:decimal("1e1")) is not allowed?
>
> Thanks,
>
> Marco.
>
Re: [basex-talk] Rounding/parsing decimal vs float
You need to use xs:float or xs:double instead of xs:decimal to be able to use
the 'e' or 'E' as the exponent separator.
Br,
Kristian K1. nov 2017 13:18 kirjutas kuupäeval Marco Lettere
:
>
> Hi all,
>
> I thought of asking this in parallel of hacking my own parsing procedure
> ...
>
> Is there a native way to parse scientific notation string into
> xs:decimal since a direct casting (xs:decimal("1e1")) is not allowed?
>
> Thanks,
>
> Marco.
>
[basex-talk] Rounding/parsing decimal vs float
Hi all,
I thought of asking this in parallel of hacking my own parsing procedure
...
Is there a native way to parse scientific notation string into
xs:decimal since a direct casting (xs:decimal("1e1")) is not allowed?
Thanks,
Marco.
Re: [basex-talk] Flexible update via function
Hi Michael,
> Manipulation of an XML structure with flexible declare function
>
> If there are some similar issues or solutions in BaseX community,
> this will be very interesting. Any other sources to learn more about?
You can have a look at the lecture slides that are listed at [1], in
particular “XQuery: More than Just a Query Language”. Our
documentation contains various practical XQuery examples as well (see
e.g. [2]). The 2nd edition of Priscilla Walmsley’s book is one of the
best available printed resources [3]. It was published end of 2015, so
it includes all new features of XQuery 3 and XQuery 3.1.
It takes a while and efforts to explore the full complexity and
variety of the XQuery language, but may save you a lot of time in the
long term.
Christian
[1] http://phobos103.inf.uni-konstanz.de/xml15
[2] http://docs.basex.org/wiki/Update
[3] http://shop.oreilly.com/product/0636920035589.do
> Christian Grün hat am 29. Oktober 2017 um 17:10 geschrieben:
>
>
> Hi Michael,
>
> Please note that an expression will always be evaluated before its
> result will be bound to a variable.
>
> You didn’t mention what result you would expect. I guess it should be
> as follows:
>
> textBBB
>
> You could pass on a path string to your function and navigate through
> the single steps, e.g. with fold-left [1]:
>
> declare function local:node-add($xml, $path, $value) {
> copy $copy := $xml
> modify (
> let $target := fold-left(
> tokenize($path, '/'),
> $copy,
> function($node, $step) {
> $node/*[name() = $step]
> }
> )
> return insert node $value into $target
> )
> return $copy
> };
>
> local:node-add(
> text,
> 'xml',
> BBB
> )
>
> If you need more flexibility, you can pass on a function that takes
> care of the navigation inside the copy/modify or update expression:
>
> declare function local:node-add($xml, $target, $value) {
> $xml update {
> insert node $value into $target(.)
> }
> };
> local:node-add(
> document { },
> function($n) { $n/A/B },
>
> )
>
> Best,
> Christian
>
> [1] http://docs.basex.org/wiki/Higher-Order_Functions#fn:fold-left
>
>
>
> On Sun, Oct 29, 2017 at 4:34 PM, wrote:
>
> Hello,
>
> How to add/replace a complete node in a node tree,
> with done by a function and by according function parameters.
> That means with flexible instructions.
>
> Such as:
>
> declare function local:NodeAdd($XML, $XMLPath, $XMLValue)
> {
> ??
> };
>
> let $XML := text
> let $XMLAdd := BBB
> let $XMLPath := $XML/xml
>
>
> return
> local:NodeAdd($XML, $XMLPath , $XMLAdd)
>
> Problem:
> Following examples didn't work, because an internal reference to
> copy-XML-structure is needed/forced.
>
> example 1: copy / modify
>
> declare function local:NodeAdd($XML, $XMLPath, $XMLValue)
> {
>
> copy $XML2 := $XML
> modify (
> insert node $XMLValue into $XML2/xml
> )
> return
> $XML2
> };
>
> But: Instead of into $XML2/xml,
>
> into $XML/xml is needed, respectively into $XMLPath
>
>
> Thanks a lot
> Michael
Re: [basex-talk] UnaryLookup expression failure
Hi Günther,
Thanks for the concise bug report. I stumbled across a similar issue
just recently [1], so I’m glad to confirm that the latest BaseX 9.0
snapshot returns the correct result [2].
All the best,
Christian
[1]
https://github.com/BaseXdb/basex/commit/ab52f2dde2628bbeac5da22898bf6e28f6826751
[2] http://files.basex.org/releases/latest/
On Mon, Oct 30, 2017 at 8:25 PM, Gunther Rademacher wrote:
> While testing some XQuery lookup expressions on BaseX 8.6.7, I got a wrong
> result for this one:
>
> (map{'a':(1,3)},map{'a':(2,4)})!?a[2]
>
> It should have returned the second item of the 'a' entry for each of the maps,
> i.e.
>
> 3 4
>
> but it returns just
>
> 3
>
> Apparently the predicate is applied as if everything before it was in
> parentheses, i.e. as if it were
>
> ((map{'a':(1,3)},map{'a':(2,4)})!?a)[2]
>
> The (simplified) parse tree for the original expression and the XQuery 3.1
> grammar however indicates that the predicate is part of the mapping target
> expression:
>
>
>
> ...
>
> !
>
>
>?
>a
>
>
>[
>2
>]
>
>
>
>
> Best regards
> Gunther
