Changeset 31269 for other-projects/hathitrust/wcsa/extracted-features-solr/trunk/solr-ingest/src/main/java/org/hathitrust/extractedfeatures/ProcessForSolrIngest.java

Timestamp:

2016-12-28T10:30:08+13:00 (7 years ago)

Author:

davidb

Message:

Some variable name changes, and printing tidy up

File:

• other-projects/hathitrust/wcsa/extracted-features-solr/trunk/solr-ingest/src/main/java/org/hathitrust/extractedfeatures/ProcessForSolrIngest.java (modified) (2 diffs)

other-projects/hathitrust/wcsa/extracted-features-solr/trunk/solr-ingest/src/main/java/org/hathitrust/extractedfeatures/ProcessForSolrIngest.java

@@ -113 +113 @@
         DoubleAccumulator progress_accum = jsc.sc().doubleAccumulator("Progress Percent");
         
-        System.err.println();
-        System.err.println();
-        System.err.println();
-        System.err.println("****##### _input_dir =  " + _input_dir);
-        System.err.println();
-        System.err.println();
-        System.err.println();
-        
         boolean icu_tokenize = Boolean.getBoolean("wcsa-ef-ingest.icu-tokenize");
         boolean strict_file_io = Boolean.getBoolean("wcsa-ef-ingest.strict-file-io");
@@ -130 +122 @@
         //json_list_data_rp.foreach(per_vol_json);
         JavaRDD<String> per_page_ids = json_list_data_rp.flatMap(per_vol_json);
-        long num_page_ids = per_page_ids.count();
-        
-        long num_ids = num_volumes;
+        long num_page_ids = per_page_ids.count(); // trigger lazy eval of: flatmap:per-vol 
+        
+        //long num_ids = num_volumes;
         
         System.out.println("");