Difference between revisions of "Lazy Module"

From BaseX Documentation
Jump to navigation Jump to search
(3 intermediate revisions by the same user not shown)
Line 24: Line 24:
 
The XQuery expression below serves as an example on how large files can be downloaded and written to a file with constant memory consumption:
 
The XQuery expression below serves as an example on how large files can be downloaded and written to a file with constant memory consumption:
  
<pre class="brush:xquery">
+
<syntaxhighlight lang="xquery">
 
file:write-binary('output.data', fetch:binary('http://files.basex.org/xml/xmark111mb.zip'))
 
file:write-binary('output.data', fetch:binary('http://files.basex.org/xml/xmark111mb.zip'))
</pre>
+
</syntaxhighlight>
  
 
If lazy items are serialized, they will be streamed as well.
 
If lazy items are serialized, they will be streamed as well.
Line 37: Line 37:
  
 
==lazy:cache==
 
==lazy:cache==
 
{{Mark|Updated with Version 9.1:}} [[#lazy:cache|lazy:cache]]: {{Code|$lazy}} argument added; support for sequences.
 
  
 
{| width='100%'
 
{| width='100%'
Line 49: Line 47:
 
* data of lazy items will be retrieved and cached inside the item.
 
* data of lazy items will be retrieved and cached inside the item.
 
* non-lazy items, or lazy items with cached data, will simply be passed through.
 
* non-lazy items, or lazy items with cached data, will simply be passed through.
* If {{Code|$lazy}} is set to {{Code|true()}}, caching will be deferred until the data is eventually requested. Streaming, however, will be disabled: Data will be cached before an input stream will be generated.
+
* If {{Code|$lazy}} is set to {{Code|true()}}, caching will be deferred until the data is eventually requested. Streaming will be disabled: Data will always be cached before a stream is returned.
 
Caching is advisable if an item will be processed more than once, or if the data may not be available anymore at a later stage.
 
Caching is advisable if an item will be processed more than once, or if the data may not be available anymore at a later stage.
 
|-
 
|-
 
| '''Example'''
 
| '''Example'''
 
|In the following example, a file will be deleted before its content is returned. To avoid a “file not found” error when serializing the result, the content must be cached:
 
|In the following example, a file will be deleted before its content is returned. To avoid a “file not found” error when serializing the result, the content must be cached:
<pre class="brush:xquery">
+
<syntaxhighlight lang="xquery">
 
let $file := 'data.txt'
 
let $file := 'data.txt'
 
let $text := lazy:cache(file:read-text($file))
 
let $text := lazy:cache(file:read-text($file))
 
return (file:delete($file), $text)
 
return (file:delete($file), $text)
</pre>
+
</syntaxhighlight>
 
|}
 
|}
  

Revision as of 15:20, 27 February 2020

This XQuery Module contains functions for handling lazy items.

In contrast to standard XQuery items, a lazy item contains a reference to the actual data, and the data itself will only be retrieved if it is requested. Hence, possible errors will be postponed, and no memory will be occupied by a lazy item as long as its content has not been requested yet.

The following BaseX functions return lazy items:

Some functions are capable of consuming the contents of lazy items in a streamable fashion: data will not be cached, but instead passed on to another target (file, the calling expression, etc.). The following streaming functions are currently available:

The XQuery expression below serves as an example on how large files can be downloaded and written to a file with constant memory consumption:

<syntaxhighlight lang="xquery"> file:write-binary('output.data', fetch:binary('http://files.basex.org/xml/xmark111mb.zip')) </syntaxhighlight>

If lazy items are serialized, they will be streamed as well.

Conventions

All functions and errors in this module are assigned to the http://basex.org/modules/lazy namespace, which is statically bound to the lazy prefix.

Functions

lazy:cache

Signatures lazy:cache($items as item()*) as item()*
lazy:cache($items as item()*, $lazy as xs:boolean) as item()*
Summary Caches the data of lazy $items in a sequence:
  • data of lazy items will be retrieved and cached inside the item.
  • non-lazy items, or lazy items with cached data, will simply be passed through.
  • If $lazy is set to true(), caching will be deferred until the data is eventually requested. Streaming will be disabled: Data will always be cached before a stream is returned.

Caching is advisable if an item will be processed more than once, or if the data may not be available anymore at a later stage.

Example In the following example, a file will be deleted before its content is returned. To avoid a “file not found” error when serializing the result, the content must be cached:

<syntaxhighlight lang="xquery"> let $file := 'data.txt' let $text := lazy:cache(file:read-text($file)) return (file:delete($file), $text) </syntaxhighlight>

lazy:is-lazy

Signatures lazy:is-lazy($item as item()) as xs:boolean
Summary Checks whether the specified $item is lazy.

lazy:is-cached

Signatures lazy:is-cached($item as item()) as xs:boolean
Summary Checks whether the contents of the specified $item are cached. The function will always return true for non-lazy items.

Changelog

Version 9.1
  • Updated: lazy:cache: $lazy argument added; support for sequences.
Version 9.0
  • Updated: Renamed from Streaming Module to Lazy Module.
  • Added: lazy:is-cached
Version 8.0

This module was introduced with Version 7.7.