MarkLogic 9 Product Documentation
xdmp:spawnxdmp:spawn(
$path as xs:string,
[$vars as item()* | map:map],
[$options as (element()|map:map)?]
) as item()*
Summary
Place the specified module on the task queue for evaluation.
| Parameters |
| path |
The path of an XQuery or JavaScript module to be executed, as a string.
The path is resolved against the root of the App Server evaluating the
query, the Modules directory, or relative to the calling module.
For details on module path resolution, see
Importing XQuery Modules, XSLT Stylesheets, and Resolving Paths in the Application Developer's Guide.
The module is considered to be JavaScript if the module path ends with
a file extension configured for the MIME type
application/vnd.marklogic-javascript in MarkLogic's
Mimetypes configuration. Otherwise, it is assumed to be XQuery.
|
| $vars |
External variable values to make available to the evaluated code,
expressed as either a sequence of alternating QName-value pairs, or
a map:map. If you use a sequence, it must contain
alternating variable QNames and values. e.g.
(xs:QName("var1"), "val1", xs:Qname("var2"), "val2").
If you use a map, then each key is a string representing the Clark
notation of the variable QName ("{namespaceURI}localname"), and its
value is the corresponding variable value. You can use
xdmp:key-from-QName
to generate the Clark notation to use as a key.
|
| options |
Options with which to customize this operation.
You can specify options as either an options XML element
in the namespace "xdmp:eval", or as a map:map. The
option names below are XML localnames. When using a map, replace any
hyphens in an option name with camel casing. For example, "an-option"
becomes "anOption" when used as a map:map entry key.
This function supports the following options:
- database
- The id of the content database. You can use functions such
as
xdmp:database
to get a database ID.
See the Usage Notes for more details. Using this option to specify
a database other than the context database requires additional
privileges. For details, see Required Privileges, below.
- modules
- The ID of the modules database to use for resolving module imports.
If you do not specify a
modules option, this operation
uses the current modules database. Use a value of 0 to
specify using the file system to resolve module imports. Using this
option may require additional privileges. For details, see Required
Privileges, below.
- root
- The root path for modules. If you do not explicitly specify a
root option, the current root is used. Use of this
option may require additional privileges. For details, see Required
Privileges, below.
- timestamp
- The system timestamp to use for this evaluation. If you omit this
option, the most recent timestamp is used. You may only specify a
timestamp for a query statement, not for an update statement. Use a
value of zero to specify the current system timestamp (the value that
would be returned by
xdmp:request-timestamp
).
For more details see
Understanding Point-In-Time Queries in the Application Developer's Guide.
Use of this option requires additional privileges. For details, see
Required Privileges, below.
- isolation
- Specify the transaction isolation for this operation. Allowed values:
same-statement, different-transaction (default).
When set to same-statement, the evaluation occurs in the
same transaction in which this function is called. When set to
different-transaction, the evaluation occurs in a separate
transaction from the one in which this function is called. If you use
same-statement isolation in a query (read-only) statement
and the eval'd code attempts an update, MarkLogic throws the exception
XDMP-UPDATEFUNCTIONFROMQUERY. For more details, see
Isolation Option to xdmp:eval/invoke in the Application Developer's Guide.
- commit
-
The commit mode for the transaction in which the code is evaluated.
Allowed values:
auto (default), explicit.
In auto mode, a transaction is committed for every statement.
In explicit mode, the transaction must be explicitely
committed or rolled back. For more details, see
Commit Mode in the Application Developer's Guide.
- update
-
Specify the transaction type in which to evaluate this code, or let
MarkLogic determine the transaction type. Allowed values:
"true", "false", "auto" (default).
For more details, see
Transaction Type Overview in the Application Developer's Guide.
- static-check
-
Whether or not to only perform a static analysis of the code, without
executing it. Allowed values:
true, false
(default).
- default-xquery-version
- The default XQuery language version to use for the query, if the query
does not contain an explicit version declaration. By default, MarkLogic
uses the default XQuery version of the App Server that called
spawn. The Task Server has no default XQuery version, so
the version is passed as part of the task request.
Allowable values for this option are "0.9-ml", "1.0-ml", "1.0", and
the special value "app-server". The first three are XQuery language
versions. The last indicates that the default XQuery language
version set on this App Server should be used. This is useful if code
written in an older XQuery version needs to call this function
on strings that may have been passed as parameters, but should be
interpreted in the App Server default language version. A module
may discover its own XQuery language version with
xdmp:xquery-version()
.
- time-limit
- Override the default time limit with this time limit for this
operation. Specify the value in seconds. You can set the value up
to the maximum-time-limit value for the App Server in which the
request is evaluated or to a lower value than the default time limit.
- user-id
- The ID of the user under which this operation should be performed.
If you do not set this option, the operation is performed as the
current user. Use of this option requires additional privileges. For
details, see Required Privileges, below. NOTE: This is a very
privileged operation since it enables a user to evaluate requests as
any other user.
- default-collation
- Specifies the collation to use for this operation if a collation
is not explicitly specified, such as in the XQuery prolog or in a
function call that allows you to specify a collation. For more
details, see
Encodings and Collations in the Search Developer's Guide.
- default-coordinate-system
- Specifies the geospatial coordinate system to use for this operation,
if a coordinate system is not explicitly specified, such as in the XQuery prolog
or in a function call that allows you to specify a coordinate system.
For more details, see
Controlling Coordinate System and Precision in the Search Developer's Guide and
Supported Coordinate Systems in the Search Developer's Guide.
- priority
- Specify the priority of the spawned task. Allowed values:
normal (default), higher.
- result
- Return a value future for the result of the spawned task.
This value future can bound be to a variable without waiting so
that work can proceed concurrently with the spawned task.
When the calling request uses the value future in any operation,
it will automatically wait for the spawned task to complete and it will
use the result. For an example, see
The second example.
- transaction-mode
- [DEPRECATED: Use the
update and commit
options instead.] Explicitly set the transaction mode for this context.
Allowed values: auto (default), query,
update-auto-commit, update. For details, see
Transaction Mode in the Application Developer's Guide.
For simple updates to be implicitly committed, specify a transaction
mode of update-auto-commit. A transaction mode of
update creates a new multi-statement update
transaction and requires an explicit commit in the code.
Within a session there can be only one active multi-statement
transaction at a time. If a new multi-statement transaction is
specified nested inside a multi-statement transaction, MarkLogic throws
the exception XDMP-NESTEDMULTI. If a new multi-statement
transaction is specified after another has been concurrently created
in the same session by another request, MarkLogic throws the exception
XDMP-SESSIONTXN and retries the current request.
An xdmp:transaction-mode XQuery prolog
option in the evaluated code overrides any transaction mode specified
with this option.
|
Required Privileges
To use this function, you must have the following privilege:
http://marklogic.com/xdmp/privileges/xdmp-spawn
To use the database option to specify a content database
other than the default database for the current App Server, you must
have the following privilege:
http://marklogic.com/xdmp/privileges/xdmp-spawn-in
To use the modules or root options to specify
to specify a modules database or root other than that configured for the
current App Server, you must have the following privilege:
http://marklogic.com/xdmp/privileges/xdmp-spawn-modules-change
To use the modules or root options to specify
to specify using the file system as the modules database or root path, you
you must have the following privilege:
http://marklogic.com/xdmp/privileges/xdmp-spawn-modules-change-file
To use the user-id
option, you must have the
following privilege:
http://marklogic.com/xdmp/privileges/xdmp-login
To use the timestamp option, you must have the following
privileges:
http://marklogic.com/xdmp/privileges/xdmp-timestamp
Usage Notes
This function places the specified module
in the task queue to be processed. The module will be evaluated when the
task server has the available resources to process it. Tasks are
processed in the order in which they are added to the queue.
Once this function is called, it cannot be rolled back,
even if the transaction from which it is called does not complete.
Therefore, use care calling this function from an update transaction.
Once a module is spawned, its evaluation is asynchronous
of the transaction in which spawn is called.
Consequently, if you call this function from a module, and if
the module ends up retrying (for example, if a deadlock is detected),
then the entire module is re-evaluated and this function is therefore
called again.
To get the database ID for options that require one, such as
the database or modules options, you can use
functions such as the following. Use the function appropriate to the
database you want to reference.
See Also
Example
xdmp:spawn("module.xqy", (),
<options xmlns="xdmp:eval">
<modules>{xdmp:modules-database()}</modules>
<root>http://example.com/application/</root>
</options>)
=> Puts the module from the modules database with the
URI http://example.com/application/module.xqy
in the task server queue.
Example
(:
This example uses the <result> option to use the results of a
spawned task in the query
:)
let $x := xdmp:spawn("/oneplusone.xqy", (),
<options xmlns="xdmp:eval">
<result>{fn:true()}</result>
</options>
)
return
($x + 2)
(:
if /oneplusone.xqy has the following body:
1 + 1
then this query returns 4
:)
Stack Overflow: Get the most useful answers to questions from the MarkLogic community, or ask your own question.