Tuesday, March 29, 2016

Oracle Endeca : Change Default indexing order in Endeca Dgraph during baseline update

Behaviour : - Records are getting indexed in Endeca and sort based on insertion order(FIFO) during baseline update from any source e.g re... thumbnail 1 summary
Behaviour : -

Records are getting indexed in Endeca and sort based on insertion order(FIFO) during baseline update from any source e.g record store, JDBC query etc. Can indexing order be changed?

Solution :- 

Dgidx process reads data from either CAS or Forge process output and stores into the index. Dgidx provides a flag to change default insertion order.

Add --sort <spec> into dgidx process. The format of <spec> is :"key|dir"

Multiple sort keys can be specified as per below  format:


"key_1[|dir_1]||key_2[|dir_2]||...||key_n[|dir_n]"

Example :- --sort in_stock|desc||best_seller|desc

Following entry needs to be added in DataIngest.xml under <Endeca_App>/config/script folder
<dgidx id="Dgidx" host-id="ITLHost">
    <properties>
      <property name="numLogBackups" value="10" />
      <property name="numIndexBackups" value="3" />
    </properties>
    <args>
      <arg>-v</arg>
      <arg>--compoundDimSearch</arg>
      <arg>--sort in_stock|desc||best_seller|desc</arg>
    </args>
    <log-dir>./logs/dgidxs/Dgidx</log-dir>
    <input-dir>./data/cas_output</input-dir>
    <output-dir>./data/dgidx_output</output-dir>
    <temp-dir>./data/temp</temp-dir>
    <run-aspell>true</run-aspell>
  </dgidx> 






No comments

Post a Comment

Text Widget