Loading the standard attribute schema

This topic describes implementation details of graphs to load the Property Description Records (PDRs) to create the standard attribute schema.

Use a CSV-based configuration graph to load the standard attribute schema. See CSV-based configuration graphs.

Input file

The PDR input file defines one or more Endeca standard attributes, with the specific settings of required and optional PDR properties. The input file resembles the following illustration:

Example input file. The file is described in detail in the text.

The first line of the sample file is the header row, which specifies the properties defined in the input file. You must specify the value for the mdex-property_Key. Other mdex-* and system navigation properties are optional. Attribute properties you do not specify either use defaults or are not defined. For information on the system default values for standard attributes, see Standard attribute default values.

In this example, the following properties are defined: 
Key,DisplayName,TextSearch,SortOrder

The names of these properties are arbitrary; you can use different names in your input file if you choose (for example, you can use AttrName instead of Key). The properties are delimited (for example, by the comma in a CSV file or the pipe character in a text file).

Note: You cannot use hyphens in the names of standard attributes. Although the data domain will accept standard attribute names that include hyphens, Integrator ETL does not. If your input includes an attribute whose name includes a hyphen, change the name in the input file. For example, if you want to input an attribute named "Sales-Type", change it in the input file to something like "Sales_Type".
The header properties map to these PDR properties:
Input Header Property Maps to PDR Property
Key mdex-property_Key
DisplayName mdex-property_DisplayName
TextSearch mdex-property_IsTextSearchable
SortOrder system-navigation_Sorting

The second and following rows in the input file contain the values for the configuration properties.

Standard attribute Transformer

In the configuration graph, use a Reformat component to build the XML to submit to the server to define your standard attribute schema. The XML consists of a set of PDRs (<mdex:record>), each of which comprises attribute properties (for example, <mdex-property_Key>, <mdex-property_DisplayName>).

The following code builds an XML from the input file illustrated above: 
integer n = 1;
integer aggrKey = 0;

// Transforms input record into output record.
function integer transform() {
  string searchBool = "";
  string saRecord = "<mdex:record xmlns=\"\">";
  saRecord = saRecord + "<mdex-property_Key>" + $0.Key + "</mdex-property_Key>";
  saRecord = saRecord + "<mdex-property_DisplayName>" + $0.DisplayName + "</mdex-property_DisplayName>";

  // Lower case the boolean in the CSV file
  searchBool = lowerCase($0.TextSearch);
  saRecord = saRecord + "<mdex-property_IsTextSearchable>" + searchBool + "</mdex-property_IsTextSearchable>";

  saRecord = saRecord + "<system-navigation_Sorting>" + $0.SortOrder + "</system-navigation_Sorting>";

  $0.xmlString = saRecord + "</mdex:record>";

  // Batch up the web service requests.
  $0.singleAggregationKey = aggrKey;
  n++;
  if (n % 15 == 0) {
    aggrKey++;
  }

  return ALL;
}

First, the code defines two integer variables, the first as a loop counter, the second as the aggregation key (aggKey).

Next, the transform() function builds the PDRs by creating the <mdex:record> container, and populating it with nodes corresponding to the properties defined in the input file (<mdex-property_Key>, <mdex-property_DisplayName>). The value of each node is derived from the corresponding value in the input file.

Note that in this example, the value of the Text Search field, which is specified in ALL CAPS, is converted to lower case before it is added to the XML.

Finally, the records are grouped into batches of 15 for submission to the server. Submitting in batches ensures a continuous stream of processing through all components in the graph, which improves performance of both Integrator ETL and the Endeca Server.

Web service request

Use the updateProperties operation of the Configuration Web Service to load the PDR configurations into the Endeca Server. This operation creates new standard attributes, and updates any affected existing standard attributes with new data. The following code illustrates a typical updateProperties operation to load PDRs: 

<config-service:configTransaction 
  xmlns:config-service="http://www.endeca.com/MDEX/config/services/types/3/0">
  <config-service:OuterTransactionId>${OUTER_TRANSACTION_ID}</config-service:OuterTransactionId>
    <config-service:updateProperties 
      xmlns:mdex="http://www.endeca.com/MDEX/XQuery/2009/09">
        $xmlString
    </config-service:updateProperties>
</config-service:configTransaction>

This example includes two variables:

  • OUTER_TRANSACTION_ID

    This variable specifies the outer transaction ID for the request. The variable and its value are stored in the workspace.prm file of the Integrator ETL project.

  • $xmlString

    This variable contains the PDRs that have been constructed by the Reformat component in the graph.

This code is added to the Web Services Client component in the configuration graph.