Enabled buffering in GPHDFS for a Parquet reader.

concourse/scripts/regression_tests_gphdfs.bash

@@ -42,6 +42,7 @@ function gen_env(){
 	tar zxf /tmp/hadoop-2.7.3.tar.gz -C /tmp
 	export HADOOP_HOME=/tmp/hadoop-2.7.3
 
+	wget -O \${HADOOP_HOME}/share/hadoop/common/lib/parquet-hadoop-bundle-1.7.0.jar http://central.maven.org/maven2/org/apache/parquet/parquet-hadoop-bundle/1.7.0/parquet-hadoop-bundle-1.7.0.jar
 	cat > "\${HADOOP_HOME}/etc/hadoop/core-site.xml" <<-EOFF
 		<configuration>
 		<property>

gpAux/extensions/gphdfs/regression/gphdfs_regress_schedule

@@ -45,3 +45,4 @@ test: varchar_Extreadwrite
 test: varchar_ExtwriteToHdfsToExtread
 test: varchar_hdfsTextToGPDB
 #test: MPP16780
+test: int_ExtreadwriteParquet
\ No newline at end of file

gpAux/extensions/gphdfs/regression/input/int_ExtreadwriteParquet.source

+-- start_ignore
+drop external table int_heap;
+drop external table int_readhdfs_parquet;
+drop external table int_writehdfs_parquet;
+-- end_ignore
+
+create readable external table int_heap(datatype_int varchar,xcount_int bigint, max_int int, min_int int, x_int int, reverse_int int, increment_int int, nullcol_int int) location ('gpfdist://%localhost%:%gpfdistPort%/int.txt')format 'TEXT';
+
+create readable external table int_readhdfs_parquet
+(
+  like int_heap
+) location ('gphdfs://%HDFSaddr%/extwrite/int_parquet') format 'PARQUET';
+-- start_ignore
+create writable external table int_writehdfs_parquet
+(
+  like int_heap
+) location ('gphdfs://%HDFSaddr%/extwrite/int_parquet') format 'PARQUET';
+-- end_ignore
+
+insert into int_writehdfs_parquet
+select * from int_heap;
+
+select count(*) from int_readhdfs_parquet;

gpAux/extensions/gphdfs/regression/output/int_ExtreadwriteParquet.source

+-- start_ignore
+drop external table int_heap;
+ERROR:  table "int_heap" does not exist
+drop external table int_readhdfs_parquet;
+ERROR:  table "int_readhdfs_parquet" does not exist
+drop external table int_writehdfs_parquet;
+ERROR:  table "int_writehdfs_parquet" does not exist
+-- end_ignore
+create readable external table int_heap(datatype_int varchar,xcount_int bigint, max_int int, min_int int, x_int int, reverse_int int, increment_int int, nullcol_int int) location ('gphdfs://localhost:9000/plaintext/int.txt')format 'TEXT';
+create readable external table int_readhdfs_parquet
+(
+  like int_heap
+) location ('gphdfs://localhost:9000/extwrite/int_parquet') format 'PARQUET';
+-- start_ignore
+create writable external table int_writehdfs_parquet
+(
+  like int_heap
+) location ('gphdfs://localhost:9000/extwrite/int_parquet') format 'PARQUET';
+NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause, defaulting to distribution columns from LIKE table
+--end_ignore
+insert into int_writehdfs_parquet
+select * from int_heap;
+select count(*) from int_readhdfs_parquet;
+ count
+-------
+  5000
+(1 row)
+
+--start_ignore
+\!/tmp/hadoop-2.7.3/bin/hadoop fs -rm -r /mapreduce/*
+rm: `/mapreduce/*': No such file or directory
+\!/tmp/hadoop-2.7.3/bin/hadoop fs -rm -r /mapred/*
+rm: `/mapred/*': No such file or directory
+--end_ignore
\ No newline at end of file

gpAux/extensions/gphdfs/src/java/1.2/com/emc/greenplum/gpdb/hadoop/formathandler/GpdbParquetFileReader.java

 package com.emc.greenplum.gpdb.hadoop.formathandler;
 
+import java.io.BufferedOutputStream;
 import java.io.DataOutputStream;
 import java.io.IOException;
 import java.io.OutputStream;
@@ -103,7 +104,7 @@ public class GpdbParquetFileReader {
 
 		Collections.sort(toReadFileList);
 
-		DataOutputStream dos = new DataOutputStream(out);
+		DataOutputStream dos = new DataOutputStream(new BufferedOutputStream(out));
 
 		int counter = 0;
 		for (FileStatus toRead : toReadFileList) {