[
Lists Home |
Date Index |
Thread Index
]
Roger L. Costello wrote:
>Hi Folks,
>
>I am wondering about the value of the "let" and "where" clause in the XQuery
>FLWOR expression. I believe that the same capability is possible simply by
>using a for-order-return expression. If that is true, then I wonder about
>the value of FLWOR (recent discussions seem to imply that FLWOR is one of
>the "key advantages" of XQuery).
>
>Let me give a couple of examples.
>
>Example 1. Let's start with a simple example. Consider this XML document:
>
><?xml version="1.0"?>
><Numbers>
> <Number>0</Number>
> <Number>8</Number>
> <Number>23</Number>
> <Number>17</Number>
> <Number>5</Number>
> <Number>19</Number>
> <Number>44</Number>
> <Number>13</Number>
> <Number>78</Number>
> <Number>21</Number>
> <Number>2</Number>
> <Number>1</Number>
> <Number>15</Number>
> <Number>67</Number>
> <Number>99</Number>
> <Number>14</Number>
> <Number>8</Number>
> <Number>33</Number>
> <Number>50</Number>
></Numbers>
>
>Problem: output all the numbers less than 20, in ascending sorted order.
>
>Here's how to do it using for-where-order-return:
>
>for $i in //Number
> where number($i) <= 20
> order by number($i) ascending
> return $i/text()
>
>
>
>
<snip/>
>Is there something that can be done with for-let-where-order-return which
>simply cannot be done with for-order-return? If not, then what value is
>for-let-where-order-return? Is it simply "syntactic sugar"? /Roger
>
>
>
for $i in //Number
let $y = number($i) * number($i) in
where number($y) <= 20
order by number($i) ascending
return $i/text()
It's possible but clumsy to achieve the same without let (imagine $y was
something more complicated). In most programming languages, "let" and
"where" are just there to make things more readable. If $y is used in
several places, an advantage can be that you evaluate it only once.
There are a number of equivalences regarding for-comprehensions and
XPath expressions - see Mary Fernandez, Jerome Simeon and Phil Wadler's
XQuery algebra paper "A semi-monad for semistructured data". In
principle, if one knows which alternative evaluates faster, one can use
these equivalences for optimization.
cheers,
Burak
--
Burak Emir
http://lamp.epfl.ch/~buraq
|
