Short Description |
Ports |
Metadata |
RunGraph Attributes |
Details |
Examples |
Compatibility |
See also |
RunGraph runs CloverETL graphs. Graph names can be specified in the component attribute or received through the input port.
Port type | Number | Required | Description | Metadata |
---|---|---|---|---|
Input | 0 | For graph names and Clover command line arguments | Input Metadata for RunGraph | |
Output | 0 | For information about graph execution | Output Metadata for RunGraph | |
1 | For information about unsuccessful graph execution | Output Metadata for RunGraph |
Information about successful execution of the specified graph is sent to the first output port.
Information about unsuccessful execution of the specified graph is sent to the second output port.
Metadata cannot be propagated through RunGraph component.
The component has metadata templates assigned to all it's ports.
Metadata on input port has to have at least 2 field.
The first two metadata fields has to be string
.
Table 61.5. Input Metadata for RunGraph
Field number | Field name | Data type | Description |
---|---|---|---|
0 | <anyname1> | string | Name of the graph to be executed, including path |
1 | <anyname2> | string | Clover command line argument.
Warning:
Arguments sent in this field are ignored
when the Same JVM attribute
is true (see RunGraph Attributes).
|
Both output ports must have the same metadata. The metadata on the output ports can have additional fields, the additional fields must placed after the metadata fields from the template.
Table 61.6. Output Metadata for RunGraph
Field number | Field name | Data type | Description |
---|---|---|---|
0 | graph | string | Name of the graph to be executed, including path |
1 | result | string | Result of graph execution
(Finished OK , Aborted ,
or Error ) |
2 | description | string | Detailed description of graph fail |
3 | message | string | Value of
org.jetel.graph.Result |
4 | duration | integer, long, or decimal | Duration of graph execution in milliseconds |
5 | runId | long | Identification of the execution of the graph which runs on CloverETL Server. |
Attribute | Req | Description | Possible values |
---|---|---|---|
Basic | |||
Graph URL | [1] | Name of one graph, including path, that should be executed by the component. Information about success or fail is sent to the first or the second output port, respectively. | |
The same JVM | By default, the same JVM instance is used to run the specified graphs.
If switched to false , graph(s)
run as external processes. When working in the server environment,
this attribute always has to be true
(thus, you cannot pass graph arguments through field 1 of port 0,
see Ports).
| true (default) | false | |
Graph parameters to pass | Parameters that are used by executed graphs.
List a sequence separated by semicolon.
If The same JVM attribute is switched to false ,
this attribute is ignored.
See
Details for more information.
| ||
Alternative JVM command | [2] | Command line to execute external process.
If you
want to give more memory to individual graphs that should be
run by this RunGraph component, type here
java -Xmx1g -cp or equivalent according to
the maximum memory needed by any of the specified
graphs. The '-cp' suffix is necessary, since classpath
is automatically appended to your command line and
the classpath attribute can differ between
various JVM implementations. | java -cp (default) | other java command |
Advanced | |||
Log file URL | Name of the file, including path, containing the log of
external processes.
The logging will be performed to the specified file independently on the value of The same JVM attribute.
If The same JVM is set true (the default setting), logging will also be performed to console.
If it is switched to false , logging to console will not be
performed and logging information will be written to the specified file.
This attribute is ignored for server environment.
See URL File Dialog.
| ||
Create directories | If set to true , non-existing directories in the
Log file URL attribute path are created. | false (default) | true | |
Append to log file | [2] | By default, data in the specified log file is overwritten on each graph run. | false (default) | true |
Graph execution class | [2] | Full class name to execute graph(s). | org.jetel.main.runGraph (default) | other execution class |
Command line arguments | [2] | Arguments of Java command to be executed when running graph(s). | |
Ignore graph fail | By default, if the execution of any of the specified graphs fails
the graph with RunGraph (that executes them) fails too.
If this attribute is set to true , fail of each executed graph is ignored.
| false (default) | true | |
[1] Either Graph URL must be specified or input port connected. [2] These attributes are applied only if The same JVM
attribute is set to |
RunGraph works in two modes. You can define graph to run using Graph URL attribute or send it using input edge.
All command line arguments passed to the RunGraph
component (either as the second field of an input record or as
the cloverCmdLineArgs
component property) are
regarded as a space delimited list of arguments which can be
quoted. Moreover, the quote character itself can be escaped by backslash.
Example 61.2. Working with Quoted Command Line Arguments
Let us have the following list of arguments:
firstArgument "second argument with spaces" "third argument with spaces and \" a quote"
The resulting command line arguments which will be passed to the child JVM are:
1) firstArgument
2) second argument with spaces
3) third argument with spaces and " a quote
Notice in 2) the argument is actually unquoted. That grants an OS-independent approach and smooth run on all platforms.
If the executed graph fails, caused error message and exception stacktrace are printed out to standard graph log. If this graph failure caused the parent graph fails (attribute Ignore graph fail is false) the caused error message is printed out on ERROR logging level, otherwise on INFO level.
RunGraph cannot run graph with subgraphs in separate JVM.
Details of graphs running in separate JVM cannot be seen in execution view of the component.
Graphs running in separate JVM does not send runId to the output.
You cannot run graph in separate JVM on CloverETL Server.
Running a graph using RunGraph |
Running several graphs using RunGraph |
Run graph ProcessInvoices.grf.
Use attribute Graph URL to define graph to be run and do not connect an input edge.
Attribute | Value |
---|---|
Graph URL | ${GRAPH_DIR}/ProcessInvoices.cdf |
Run several graphs using RunGraph component.
Send graph URLs to the RunGraph using input edge. Do not fill in the attribute Graph URL.
Previous version of CloverETL required to specify Command line arguments if path to graph is set up using Graph URL attribute.