An update adapter is part of a partial update pipeline. It writes partial update information to a running Endeca MDEX Engine to perform a live update.
The partial update pipeline reads its input as normal records, using a normal record adapter. The input records are transformed into record updates using the update record expression, then written out using an update adapter. An update adapter is typically the last component of a partial update pipeline.
Forge processing for full updates and partial updates is done in separate pipelines. Therefore, you would never add an update adapter to a full update pipeline. See the Endeca Partial Updates Guide for detailed information on how to configure a partial update pipeline.
To add an update adapter:
In the Pipeline Diagram editor, click New, and then choose Update Adapter.
The Update Adapter editor appears.
In the Name text box, type a unique name for this update adapter.
In the General tab, do the following:
In the Output URL text box, type the location to which Forge writes its files and processed records.
( Optional) In the Output prefix text box, type a filename prefix to add when Forge writes its output files. If none is specified, Forge uses the prefix
Out.(Optional) Check Multi if you want Forge to be able to read from multiple files.
If you checked Multi in step C , then in the Multi property text box, type a property name that indicates the location to which a record coming through Forge should be written
If you enabled Filter unknown properties in your full update pipeline's indexer adapter, enable Filter unknown properties in the update adapter.
(Optional) If you are using an Agraph in your implementation, do the following:
If you want to change the partition property, open the Properties view and modify which properties are enabled for rollup and record spec. For more detailed information, see the Endeca Advanced Development Guide.
(Optional) In the Comment tab, add a comment for the component.
Fully implementing this feature requires additional work outside of Developer Studio. Please refer to the Endeca Partial Updates Guide for details.
The Update Adapter editor contains a unique name for this update adapter.
The Update Adapter editor contains the following tabs:
The General tab contains the following options:
|
Option |
Description |
|---|---|
|
Output URL |
Required. Identifies the directory to which Forge writes its files and processed records. The path is either an absolute path or a path relative to the location of the Pipeline.epx file. With an absolute path, the protocol must be specified in RFC 2396 syntax. Usually this means the prefix file:/// precedes the path to the data file. Relative URLs must not specify the protocol and are relative to the URL used to locate the Pipeline.epx file. |
|
Output prefix |
Specifies the filename prefix to add when Forge writes its output files. |
|
Multi |
Optional. The Multi and Multi property fields specify which file each record is written to. If Multi is checked, Forge checks the Multi property field for the location. |
|
Multi property |
Optional . Specifies a property name that indicates the location that a record coming through Forge should be written to. (Forge stores a property on each record that is the source URL for that record.) Setting this property is useful if you want to process multiple input files and also keep the output distinct. |
|
Filter unknown properties |
Optional. If you enabled Filter unknown properties in your full update pipeline's indexer adapter , enable Filter unknown properties in the update adapter. |
|
Custom Compression level |
Optional. Sets the level of compression to be performed on the data. The values can range from 0 to 10, with higher numbers indicating higher compression (smaller size, slower processing). |
The Sources tab contains the following options:
The Agraph tab contains the following options:
|
Option |
Description |
|---|---|
|
Enable Agraph Support |
When checked, enables the Agraph program for use in a partial update pipeline. |
|
Number of Agraph partitions |
Specifies the number of child Dgraphs that the Agraph controls. In an Agraph implementation, this must be a value of 2 or more. |
|
Partition property |
The partition property is a read only field that identifies the property by which records are assigned to each partition. If you are using rollup capabilities, the rollup property displays as the partition property. If you do not have a roll up property, but do have a record spec property enabled in your project, the record spec property functions as the partition property. If neither a rollup nor a record spec property exists, the partition property is empty, and Forge assigns records to each partition according to a round-robin strategy. |
See the Endeca Advanced Development Guide for detailed information on how to configure, provision, and run an Aggregated MDEX Engine implementation.
