Added support for fieldParent

Trax-air · Jul 31, 2014 · 1a80d84 · 1a80d84
1 parent 8c5c65e
commit 1a80d84
Show file tree

Hide file tree

Showing 5 changed files with 20 additions and 0 deletions.
diff --git a/README.md b/README.md
@@ -89,6 +89,7 @@ The CSV river import data from CSV files and index it.
 	        "escape_character" : ";",
 	        "quote_character" : "'",
             "field_id" : "id",
+            "field_parent" : "id_parent",
             "concurrent_requests" : "1",
             "charset" : "UTF-8",
             "script_before_all": "/path/to/before_all.sh",

diff --git a/src/main/groovy/org/agileworks/elasticsearch/river/csv/Configuration.groovy b/src/main/groovy/org/agileworks/elasticsearch/river/csv/Configuration.groovy
@@ -32,6 +32,7 @@ class Configuration {
     int bulkThreshold
     int concurrentRequests
     String idField
+    String idParent
 
     String scriptBeforeAll
     String scriptAfterAll
@@ -55,6 +56,7 @@ class Configuration {
             separator = nodeStringValue(csvSettings.get(Constants.CSV.FIELD_SEPARATOR), String.valueOf(',')).charAt(0)
             quoteCharacter = nodeStringValue(csvSettings.get(Constants.CSV.QUOTE_CHARACTER), String.valueOf('\"')).charAt(0)
             idField = nodeStringValue(csvSettings.get(Constants.CSV.FIELD_ID), 'id')
+            parentField = nodeStringValue(csvSettings.get(Constants.CSV.FIELD_PARENT), '')
             concurrentRequests = nodeIntegerValue(csvSettings.get(Constants.CSV.CONCURRENT_REQUESTS), 1)
 
             String charsetName = nodeStringValue(csvSettings.get(Constants.CSV.CHARSET), 'UTF-8')

diff --git a/src/main/groovy/org/agileworks/elasticsearch/river/csv/Constants.groovy b/src/main/groovy/org/agileworks/elasticsearch/river/csv/Constants.groovy
@@ -26,6 +26,7 @@ class Constants {
         static final String FIELD_SEPARATOR = 'field_separator'
         static final String QUOTE_CHARACTER = 'quote_character'
         static final String FIELD_ID = 'field_id'
+        statif final String FIELD_PARENT = 'field_parent'
 
         static final String CHARSET = 'charset'
 

diff --git a/src/main/groovy/org/agileworks/elasticsearch/river/csv/OpenCSVFileProcessor.groovy b/src/main/groovy/org/agileworks/elasticsearch/river/csv/OpenCSVFileProcessor.groovy
@@ -77,6 +77,10 @@ class OpenCSVFileProcessor implements FileProcessor {
             request.id(UUID.randomUUID().toString())
         }
 
+        if (csvContainsParentColumn()) {
+            request.parent(getParent(line))
+        }
+
         request.create(false).source(builder)
 
         listener.onLineProcessed(request)
@@ -86,10 +90,21 @@ class OpenCSVFileProcessor implements FileProcessor {
         return config.csvFields.find { it == config.idField }
     }
 
+    boolean csvContainsParentColumn() {
+        return config.csvFields.find { it == config.idParent }
+    }
+
     String getId(String[] line) {
 
         int index = config.csvFields.indexOf(config.idField)
 
         return line[index]
     }
+
+    String getParent(String[] line) {
+
+        int index = config.csvFields.indexOf(config.idParent)
+
+        return line[index]
+    }
 }
diff --git a/src/test/groovy/org/agileworks/elasticsearch/river/csv/ConfigurationTest.groovy b/src/test/groovy/org/agileworks/elasticsearch/river/csv/ConfigurationTest.groovy
@@ -79,6 +79,7 @@ class ConfigurationTest extends Specification {
         config.filenamePattern
         !config.folderName
         config.idField
+        config.idParent
         config.indexName
         config.poll
         config.quoteCharacter