Merge pull request #37 from cloudsufi/snowflake-escape-char-conf

[PLUGIN-1816] Added fix for decimal issue not having rounding mode and made escape character configurable.

Author: vikasrathee-cs

src/main/java/io/cdap/plugin/snowflake/common/util/SchemaHelper.java

@@ -24,6 +24,7 @@
 import io.cdap.plugin.snowflake.common.client.SnowflakeFieldDescriptor;
 import io.cdap.plugin.snowflake.common.exception.SchemaParseException;
 import io.cdap.plugin.snowflake.source.batch.SnowflakeBatchSourceConfig;
+import io.cdap.plugin.snowflake.source.batch.SnowflakeInputFormatProvider;
 import io.cdap.plugin.snowflake.source.batch.SnowflakeSourceAccessor;
 import java.io.IOException;
 import java.sql.Types;
@@ -62,7 +63,8 @@ public static Schema getSchema(SnowflakeBatchSourceConfig config, FailureCollect
       return getParsedSchema(config.getSchema());
     }
 
-    SnowflakeSourceAccessor snowflakeSourceAccessor = new SnowflakeSourceAccessor(config);
+    SnowflakeSourceAccessor snowflakeSourceAccessor =
+      new SnowflakeSourceAccessor(config, SnowflakeInputFormatProvider.PROPERTY_DEFAULT_ESCAPE_CHAR);
     return getSchema(snowflakeSourceAccessor, config.getSchema(), collector, config.getImportQuery());
   }
 

src/main/java/io/cdap/plugin/snowflake/source/batch/SnowflakeBatchSource.java

@@ -16,6 +16,7 @@
 
 package io.cdap.plugin.snowflake.source.batch;
 
+import com.google.common.base.Strings;
 import io.cdap.cdap.api.annotation.Description;
 import io.cdap.cdap.api.annotation.Name;
 import io.cdap.cdap.api.annotation.Plugin;
@@ -33,6 +34,7 @@
 import io.cdap.plugin.snowflake.common.util.SchemaHelper;
 import org.apache.hadoop.io.NullWritable;
 
+import java.util.HashMap;
 import java.util.Map;
 import java.util.stream.Collectors;
 
@@ -68,7 +70,11 @@ public void configurePipeline(PipelineConfigurer pipelineConfigurer) {
   public void prepareRun(BatchSourceContext context) {
     FailureCollector failureCollector = context.getFailureCollector();
     config.validate(failureCollector);
-
+    Map<String, String> arguments = new HashMap<>(context.getArguments().asMap());
+    String escapeChar = arguments.containsKey(SnowflakeInputFormatProvider.PROPERTY_ESCAPE_CHAR) &&
+      !Strings.isNullOrEmpty(arguments.get(SnowflakeInputFormatProvider.PROPERTY_ESCAPE_CHAR))
+      ? arguments.get(SnowflakeInputFormatProvider.PROPERTY_ESCAPE_CHAR)
+      : SnowflakeInputFormatProvider.PROPERTY_DEFAULT_ESCAPE_CHAR;
     Schema schema = SchemaHelper.getSchema(config, failureCollector);
     failureCollector.getOrThrowException();
 
@@ -81,7 +87,7 @@ public void prepareRun(BatchSourceContext context) {
                                    .collect(Collectors.toList()));
     }
 
-    context.setInput(Input.of(config.getReferenceName(), new SnowflakeInputFormatProvider(config)));
+    context.setInput(Input.of(config.getReferenceName(), new SnowflakeInputFormatProvider(config, escapeChar)));
   }
 
   @Override

src/main/java/io/cdap/plugin/snowflake/source/batch/SnowflakeInputFormat.java

@@ -60,6 +60,8 @@ private SnowflakeSourceAccessor getSnowflakeAccessor(Configuration configuration
       SnowflakeInputFormatProvider.PROPERTY_CONFIG_JSON);
     SnowflakeBatchSourceConfig config = GSON.fromJson(
       configJson, SnowflakeBatchSourceConfig.class);
-    return new SnowflakeSourceAccessor(config);
+    String escapeChar = configuration.get(SnowflakeInputFormatProvider.PROPERTY_ESCAPE_CHAR,
+                                          SnowflakeInputFormatProvider.PROPERTY_DEFAULT_ESCAPE_CHAR);
+    return new SnowflakeSourceAccessor(config, escapeChar);
   }
 }

src/main/java/io/cdap/plugin/snowflake/source/batch/SnowflakeInputFormatProvider.java

@@ -29,13 +29,17 @@
 public class SnowflakeInputFormatProvider implements InputFormatProvider {
 
   public static final String PROPERTY_CONFIG_JSON = "cdap.snowflake.source.config";
+  public static final String PROPERTY_ESCAPE_CHAR = "cdap.snowflake.source.escape";
+
+  public static final String PROPERTY_DEFAULT_ESCAPE_CHAR = "\\";
 
   private static final Gson GSON = new Gson();
   private final Map<String, String> conf;
 
-  public SnowflakeInputFormatProvider(SnowflakeBatchSourceConfig config) {
+  public SnowflakeInputFormatProvider(SnowflakeBatchSourceConfig config, String escapeChar) {
     this.conf = new ImmutableMap.Builder<String, String>()
       .put(PROPERTY_CONFIG_JSON, GSON.toJson(config))
+      .put(PROPERTY_ESCAPE_CHAR, escapeChar)
       .build();
   }
 

src/main/java/io/cdap/plugin/snowflake/source/batch/SnowflakeMapToRecordTransformer.java

@@ -24,6 +24,7 @@
 import org.slf4j.LoggerFactory;
 
 import java.math.BigDecimal;
+import java.math.RoundingMode;
 import java.time.Instant;
 import java.time.LocalDate;
 import java.time.LocalTime;
@@ -83,7 +84,8 @@ private Object convertValue(String fieldName, String value, Schema fieldSchema)
         case TIME_MICROS:
           return TimeUnit.NANOSECONDS.toMicros(LocalTime.parse(value).toNanoOfDay());
         case DECIMAL:
-          return new BigDecimal(value).setScale(fieldSchema.getScale()).unscaledValue().toByteArray();
+            return new BigDecimal(value).setScale(fieldSchema.getScale(),
+                                                  RoundingMode.HALF_EVEN).unscaledValue().toByteArray();
         default:
           throw new IllegalArgumentException(
             String.format("Field '%s' is of unsupported type '%s'", fieldSchema.getDisplayName(),

src/main/java/io/cdap/plugin/snowflake/source/batch/SnowflakeSourceAccessor.java

@@ -60,10 +60,12 @@ public class SnowflakeSourceAccessor extends SnowflakeAccessor {
       "OVERWRITE=TRUE HEADER=TRUE SINGLE=FALSE";
   private static final String COMMAND_MAX_FILE_SIZE = " MAX_FILE_SIZE=%s";
   private final SnowflakeBatchSourceConfig config;
+  private final char escapeChar;
 
-  public SnowflakeSourceAccessor(SnowflakeBatchSourceConfig config) {
+  public SnowflakeSourceAccessor(SnowflakeBatchSourceConfig config, String escapeChar) {
     super(config);
     this.config = config;
+    this.escapeChar = escapeChar.charAt(0);
   }
 
   /**
@@ -116,7 +118,7 @@ public CSVReader buildCsvReader(String stageSplit) throws IOException {
       InputStream downloadStream = connection.unwrap(SnowflakeConnection.class)
         .downloadStream("@~", stageSplit, true);
       InputStreamReader inputStreamReader = new InputStreamReader(downloadStream);
-      return new CSVReader(inputStreamReader);
+      return new CSVReader(inputStreamReader, ',', '"', escapeChar);
     } catch (SQLException e) {
       throw new IOException(e);
     }

src/test/java/io/cdap/plugin/snowflake/common/client/SnowflakeAccessorTest.java

@@ -18,6 +18,7 @@
 
 import io.cdap.plugin.snowflake.Constants;
 import io.cdap.plugin.snowflake.common.BaseSnowflakeTest;
+import io.cdap.plugin.snowflake.source.batch.SnowflakeInputFormatProvider;
 import io.cdap.plugin.snowflake.source.batch.SnowflakeSourceAccessor;
 import org.junit.Assert;
 import org.junit.Test;
@@ -44,7 +45,8 @@
  */
 public class SnowflakeAccessorTest extends BaseSnowflakeTest {
 
-  private SnowflakeSourceAccessor snowflakeAccessor = new SnowflakeSourceAccessor(CONFIG);
+  private SnowflakeSourceAccessor snowflakeAccessor =
+    new SnowflakeSourceAccessor(CONFIG, SnowflakeInputFormatProvider.PROPERTY_DEFAULT_ESCAPE_CHAR);
 
   @Test
   public void testDescribeQuery() throws Exception {