Re: [Videlibri-xidel] How to merge preceding empty nodes?

[Benito v. d. Z. <be...@be...>]

Hi Peng,

the standard has a few examples: 
https://www.w3.org/TR/xquery-31/#id-tumbling-windows

You can also ask ChatGPT about these things:
Explain the tumbling window syntax of XQuery.

In XQuery, a tumbling window is a construct used to perform calculations 
on a set of consecutive items within a sequence. The syntax for defining 
a tumbling window in XQuery involves the use of the "window" clause, 
which allows you to specify the size and sliding behavior of the window.

Here is the syntax for the tumbling window clause in XQuery:

sqlCopy code
|for tumbling window $window in $sequence start $start when $condition 
end $end when $condition return $result |

Let's break down each part of the syntax:

  * "for tumbling window": This specifies that you are creating a
    tumbling window.
  * "$window": This is the name of the window that you are defining.
  * "in $sequence": This specifies the sequence of items that you want
    to operate on.
  * "start $start when $condition": This specifies when the window
    should start. "$start" is the name of the variable that holds the
    starting item, and "$condition" is a boolean expression that
    evaluates to true when the starting item is found.
  * "end $end when $condition": This specifies when the window should
    end. "$end" is the name of the variable that holds the ending item,
    and "$condition" is a boolean expression that evaluates to true when
    the ending item is found.
  * "return $result": This specifies the result of the window operation.
    "$result" is an expression that calculates the result based on the
    items in the window.

In essence, the tumbling window syntax allows you to define a sliding 
window that "tumbles" over the sequence of items, performing some 
operation on each set of consecutive items within the window.



Given an example that shows how to use this syntax with nodes from an 
XML document.

Sure! Let's say we have an XML document containing information about 
books, like this:

xmlCopy code
|<books>   <book>     <title>Harry Potter and the Philosopher's 
Stone</title>     <author>J.K. Rowling</author>     <year>1997</year>   
</book>   <book>     <title>The Hobbit</title>     <author>J.R.R. 
Tolkien</author>     <year>1937</year>   </book>   <book>     <title>To 
Kill a Mockingbird</title>     <author>Harper Lee</author>     
<year>1960</year>   </book>   <book>     <title>The Catcher in the 
Rye</title>     <author>J.D. Salinger</author>     <year>1951</year>   
</book> </books> |

And we want to calculate the average publication year of books in sets 
of three, starting from the first book.

We can use the tumbling window syntax to accomplish this:

xqueryCopy code
|let $books := doc("books.xml")/books/book for tumbling window $window 
in $books     start at $start when position() = 1     end at $end when 
position() mod 3 = 0     return avg($window/year) |

In this example, we first assign the sequence of "book" nodes to the 
variable "$books". Then, we create a tumbling window with the "for 
tumbling window" clause, using the variable "$books" as the sequence.

The "start at" clause specifies that the window should start at the 
first item in the sequence, using the "position()" function to test the 
position of the current item.

The "end at" clause specifies that the window should end at every third 
item in the sequence, again using the "position()" function and the 
modulo operator to check if the current position is a multiple of three.

Finally, we use the "return" clause to calculate the average publication 
year of the items in each window, using the "avg()" function to average 
the "year" elements of the window.

When we execute this query, we get the following result:

xmlCopy code
|19.57 1951 |

The first line of the result is the average publication year of the 
first three books, and the second line is the average publication year 
of the last book.


Bye,
Benito
On 22.04.23 12:09, Peng Yu wrote:
> I am not familiar with tumbling window. It is not clear to me where is
> a good resource for learning it quickly. Do you have a good reference?
>
> On 4/21/23, Benito van der Zander<be...@be...>  wrote:
>> Hi Peng,
>>
>> XQuery got an amazingly complicated syntax for that:
>>
>> xidel /tmp/x.xml --xquery '<a>{for tumbling window $w in /a/x
>>                                  start when true()
>>                                  end $x when $x != ""
>>                                  return <x
>> label="{string-join($w/@label)}">{$x/node()}</x>
>>                              }</a> ' --xml
>>
>>
>>
>> Cheers,
>> Benito
>> On 19.04.23 06:08, Peng Yu wrote:
>>> <a>
>>> <x label="1"></x>
>>> <x label="2"></x>
>>> <x label="3">y</x>
>>> <x label="4"></x>
>>> <x label="5"></x>
>>> <x label="6">z</x>
>>> </a>
>>>
>>> For the above input, I want to merge the preceding empty nodes with
>>> nonempty ones.
>>>
>>> The output should look like this.
>>>
>>> <a>
>>> <x label="123">y</x>
>>> <x label="456">z</x>
>>> </a>
>>>
>>> How can I make this transform with xidel?
>>>
>