Changeset 28794 for main/trunk/model-sites-dev/von-sparql/collect/nz-natlib-cat

Timestamp:

2014-01-15T14:35:55+13:00 (10 years ago)

Author:

ak19

Message:

Encoding issue now straightened out

File:

• main/trunk/model-sites-dev/von-sparql/collect/nz-natlib-cat/pre-import/marcXML_Split/src/split.java (modified) (3 diffs)

main/trunk/model-sites-dev/von-sparql/collect/nz-natlib-cat/pre-import/marcXML_Split/src/split.java

@@ -13 +13 @@
         //Default values for arguments
         int NRecords = 250;
-        String OutputPath = "./out/";
+        String OutputPath = "./out/nz_out";
         String InputPath = "./NZDataFull.xml";
+        //String InputPath = "./NZ-small-macron-test.xml";
 
         //Incorrect number of arguments supplied
-        if(args.length!=4)
+        if(args.length!=2)
         {
-            System.err.println("USAGE: java split [-n records_per_file] [-o output_path]");
+            System.err.println("USAGE: ./RUN [records_per_file] [output_path]");
             return;
         }
-        //read arguments    
-        for(int i=0; i < args.length; i+=2)
+        
+        try{
+            NRecords = Integer.parseInt(args[0]);
+            OutputPath = args[1];
+        }
+        catch(Exception e)
         {
-            if(args[i].equals("-n"))
-                NRecords = Integer.parseInt(args[i+1]);
-            else if(args[i].equals("-o"))
-                OutputPath = args[i+1];
-        }
-        
+            System.err.println("Invalid Arguments");
+            System.err.println("USAGE: ./RUN [records_per_file] [output_path]");
+            return;
+        }   
+
         InputStream in;
         try{
@@ -44 +48 @@
         File f;
         int RecordCount = 0;
+                int ParseErrorCount = 0;
         
         while(reader.hasNext())
-        {       
+        {   
             f = new File(TempFilename);
-            MarcWriter writer = new MarcXmlWriter(new FileOutputStream(f),true);
+            MarcWriter writer = new MarcXmlWriter(new FileOutputStream(f),"UTF8",true);
             Record record;
 
@@ -54 +59 @@
             for(int i=0; (i<NRecords && reader.hasNext()); i++)
             {
-                record = reader.next();
-                writer.write(record);
-                RecordCount++;
+                try {
+                    record = reader.next();
+                }
+                catch (Exception e) {
+                    System.err.println("Warning: Failed to parse record number: " + RecordCount);
+                    ParseErrorCount++;
+                    continue;
+                }
+                                writer.write(record);
+                                RecordCount++;
             }
             writer.close();