[xquery-talk] Function and Query Evaluation with No XML Tags Error

[Wei, Alice J.]

Hi, Kevin:

   Thanks, this does bring some help to what I am trying to work on now. To certain extent, I have this feeling that the syntax between XQuery functions and XQuery user-defined functions are somewhat different.

  Your code does point out quite a few different options and the variety of errors I could be getting when I am not doing them correctly. It does do the trick.

Thanks for your help.

Alice
======================================================
Alice Wei
MIS 2008
School of Library and Information Science
Indiana University Bloomington
ajwei at indiana.edu
________________________________________
From: kogrover at gmail.com [kogrover at gmail.com] On Behalf Of Kevin Grover [kevin at kevingrover.net]
Sent: Tuesday, February 26, 2008 9:38 PM
To: Wei, Alice J.
Cc: talk at x-query.com
Subject: Re: [xquery-talk] Function and Query Evaluation with No XML Tags Error

On Tue, Feb 26, 2008 at 8:21 AM, Wei, Alice J. <ajwei at indiana.edu<mailto:ajwei at indiana.edu>> wrote:
Hi, David:

 I think I am having a different issue different from what I had several weeks ago.
  What I have modified is from:

let $ad := fn:collection("xmldb:exist://db/cbml")//ad/p[contains(upper-case(.), 'BOOK')]
let $sorted_result:=
for $doc in distinct-values($ad)
order by $doc
return $doc
for $r at $count in $sorted_result
let $nodes := $ad[. = $r][1]
return
<ad>
{$nodes}
</ad>

 by trying to put it in user-defined functions, which is currently giving me quite a bit of problems from itself.

declare boundary-space preserve;
declare function local:distinct(
    $seq as xs:anyAtomicType*)
     as xs:anyAtomicType
 {
 let $doc := distinct-values($seq)
 order by $doc
 return
 <ad>{$doc} </ad>
 };

 for $ad in
distinct-values(collection("xmldb:exist://db/cbml")//ad/p[contains(upper-case(.$
 return
<ad>{local:distinct($ad)}</ad>

As I mentioned before, I only intend to have the output be appeared in the same way as the above XQuery that is not in the function.
Is this possible?

Thanks for your help.
======================================================
Alice Wei
MIS 2008
School of Library and Information Science
Indiana University Bloomington
ajwei at indiana.edu<mailto:ajwei at indiana.edu>
________________________________________
From: David Carlisle [davidc at nag.co.uk<mailto:davidc at nag.co.uk>]
Sent: Tuesday, February 26, 2008 10:33 AM
To: Wei, Alice J.
Cc: talk at x-query.com<mailto:talk at x-query.com>
Subject: Re: [xquery-talk] Function and Query Evaluation with No XML Tags Error

Alice,

Alice it seems hard to help as you have been essentially posting the
same code, with the same error for weeks,

When you go

 for $ad in distinct-values(...anything)

then, on each iteration $ad will be a single string (one of the values)

So on each call to local:distinct($ad) you are passing in a single string. so
doing operations such as distinct-values() or for or order by are all
essentialy null-operations. It's not an error to take all the
unique values from a sequence of length one and then sort them, but you will always
just get back the value you started with.


See for example:

http://x-query.com/pipermail/talk/2008-January/002436.html

David


distinct-values returns a list of values (strings) NOT elements.  (hence the 'values').   It's throwing away all of the XML element markup and just keeping the string values (all smashed together).

Example XML file:
<?xml version="1.0"?>
<data>
  <record>
    <name>Fred Flinstone</name>
    <street>Fred's Street</street>
  </record>
  <record>
    <name>Fred Flinstone</name>
    <street>Fred's Street</street>
  </record>
  <record>
    <name>Barney Rubble</name>
    <street>Barney's Street</street>
  </record>
  <record>
    <name>Wilma Flinstone</name>
    <street>Wilma's Street</street>
  </record>
  <record>
    <name>Betty Rubble</name>
    <street>Betty's Street</street>
  </record>
</data>

NOTICE: Fred's entry is duplicated!

Example Query:
for $r in distinct-values(/data/record)
return $r


Results:
<?xml version="1.0" encoding="UTF-8"?>
    Fred Flinstone
    Fred's Street

    Barney Rubble
    Barney's Street

    Wilma Flinstone
    Wilma's Street

    Betty Rubble
    Betty's Street

NOTICE: No XML markup.  It was all stripped.  It's like calling string on it: <record><name> and <street> have been removed: the data flattened into a string.

It's like this:

for $r in /data/record
return string($r)


Results:
<?xml version="1.0" encoding="UTF-8"?>
    Fred Flinstone
    Fred's Street

    Fred Flinstone
    Fred's Street

    Barney Rubble
    Barney's Street

    Wilma Flinstone
    Wilma's Street

    Betty Rubble
    Betty's Street

Again: no XML markup.  They're strings (because I explicitly converted them to strings).  However, this time, I DID get the duplicated entry for Fred.

Will you really have different nodes (with subnodes) that contain completely duplicated data?  If so, I have no suggestions.  If not and you're really getting multiple instances of the SAME node, then use the id function to test for uniqueness.

I figured I could use this:
(: ** Generates runtime errror ** :)
for $r in /data/record[not(string(.)=string(preceding::*))]
return string($r)

But it generates a complaint (at runtime):
Description: A sequence of more than one item is not allowed as the first argument of string() (<record/>, <name/>, ...)
** SEE More at the bottom: I figured it out.

I tried a few other things to get distinct nodes based on the content, but none of the others would even parse : I'm still mostly in the dark about the magic of XQuery and XPath.

Try removing the disctinct-values (and iterate over the elements) and you may get something closer to what you are expecting.

** (corrected from above)
Here's a unique by value
for $r in /data/record[not(string(.)=string((preceding-sibling::*)[1]))]
return string($r)

Results:
<?xml version="1.0" encoding="UTF-8"?>
    Fred Flinstone
    Fred's Street

    Barney Rubble
    Barney's Street

    Wilma Flinstone
    Wilma's Street

    Betty Rubble
    Betty's Street

Again.  It's unique (because I made it unique in the XPath), but it's iterating over the nodes.  It's a string, because I explicitly convert them to strings.  I could have also done:

for $r in /data/record[not(string(.)=string((preceding-sibling::*)[1]))]
return normalize-space($r)


Results:
<?xml version="1.0" encoding="UTF-8"?>Fred Flinstone Fred's Street Barney Rubble Barney's Street Wilma Flinstone Wilma's Street Betty Rubble Betty's Street

NOTICES: it all a string (because normalize-space converts things to string) but 'extra' whitespace is stripped.

Or:
for $r in /data/record[not(string(.)=string((preceding-sibling::*)[1]))]
return <newelement>{normalize-space($r)}</newelement>

Results:
<?xml version="1.0" encoding="UTF-8"?>
<newelement>Fred Flinstone Fred's Street</newelement>
<newelement>Barney Rubble Barney's Street</newelement>
<newelement>Wilma Flinstone Wilma's Street</newelement>
<newelement>Betty Rubble Betty's Street</newelement>

NOTICE: String results wrapped in a new tag.

Or:
for $r in /data/record[not(string(.)=string((preceding-sibling::*)[1]))]
return <newelement>{$r}</newelement>

Results:
<?xml version="1.0" encoding="UTF-8"?>
<newelement>
   <record>
      <name>Fred Flinstone</name>
      <street>Fred's Street</street>
  </record>
</newelement>
<newelement>
   <record>
      <name>Barney Rubble</name>
      <street>Barney's Street</street>
  </record>
</newelement>
<newelement>
   <record>
      <name>Wilma Flinstone</name>
      <street>Wilma's Street</street>
  </record>
</newelement>
<newelement>
   <record>
      <name>Betty Rubble</name>
      <street>Betty's Street</street>
  </record>
</newelement>

NOTICE: Original XML elements (and sub elements) wrapped in a new tag.

Or, if you plan on using that 'logic' over and over, you can create a function for it:

XQuery: This is the same as (functionally) the previous example, it just defines (and uses) a local function

declare function local:unique-nodes-by-value($seq as element()*) as element()*
{
  for $r in $seq[not(string(.)=string((preceding-sibling::*)[1]))]
  return $r
};

for $r in local:unique-nodes-by-value(/data/record)
return <newelement>{$r}</newelement>

Using XPaths, you can pick and choose what element/sub element (values, content, or sub-content) you get.

I hope this helps some.

Keep in mind, I'm relatively new at this so there are probably better ways.

- Kevin