Chetan Bhat created CARBONDATA-3987:
---------------------------------------
Summary: Issues in SDK Pagination reader (2 issues)
Key: CARBONDATA-3987
URL: https://issues.apache.org/jira/browse/CARBONDATA-3987
Project: CarbonData
Issue Type: Bug
Components: other
Affects Versions: 2.1.0
Reporter: Chetan Bhat
Issue 1 :
write data to table and insert into one more row , error is thrown when try to
read new added row where as getTotalRows get incremented by 1.
Test code-
/**
* Carbon Files are written using CarbonWriter in outputpath
*
* Carbon Files are read using paginationCarbonReader object
* Checking pagination with insert on large data with 8 split
*/
@Test
public void testSDKPaginationInsertData() throws IOException,
InvalidLoadOptionException, InterruptedException {
System.out.println("___________________________________________" +
name.getMethodName() + " TestCase Execution is
started________________________________________________");
//
// String outputPath1 = getOutputPath(outputDir, name.getMethodName() +
"large");
//
// long uid = 123456;
// TimeZone.setDefault(TimeZone.getTimeZone("Asia/Shanghai"));
// writeMultipleCarbonFiles("id int,name string,rank short,salary double,active
boolean,dob date,doj timestamp,city string,dept string", getDatas(),
outputPath1, uid, null, null);
//
// System.out.println("Data is written");
List<String[]> data1 = new ArrayList<String[]>();
String[] row1 = \{"1", "AAA", "3", "3444345.66", "true", "1979-12-09",
"2011-2-10 1:00:20", "Pune", "IT"};
String[] row2 = \{"2", "BBB", "2", "543124.66", "false", "1987-2-19",
"2017-1-1 12:00:20", "Bangalore", "DATA"};
String[] row3 = \{"3", "CCC", "1", "787878.888", "false", "1982-05-12",
"2015-12-1 2:20:20", "Pune", "DATA"};
String[] row4 = \{"4", "DDD", "1", "99999.24", "true", "1981-04-09",
"2000-1-15 7:00:20", "Delhi", "MAINS"};
String[] row5 = \{"5", "EEE", "3", "545656.99", "true", "1987-12-09",
"2017-11-25 04:00:20", "Delhi", "IT"};
data1.add(row1);
data1.add(row2);
data1.add(row3);
data1.add(row4);
data1.add(row5);
String outputPath1 = getOutputPath(outputDir, name.getMethodName() + "large");
long uid = 123456;
TimeZone.setDefault(TimeZone.getTimeZone("Asia/Shanghai"));
writeMultipleCarbonFiles("id int,name string,rank short,salary double,active
boolean,dob date,doj timestamp,city string,dept string", data1, outputPath1,
uid, null, null);
System.out.println("Data is written");
String hdfsPath1 = moveFiles(outputPath1, outputPath1);
String datapath1 = hdfsPath1.concat("/" + name.getMethodName() + "large");
System.out.println("HDFS Data Path is: " + datapath1);
runSQL("create table " + name.getMethodName() + "large" + " using carbon
location '" + datapath1 + "'");
System.out.println("Table " + name.getMethodName() + " is created
Successfully");
runSQL("select count(*) from " + name.getMethodName() + "large");
long uid1 = 123;
String outputPath = getOutputPath(outputDir, name.getMethodName());
List<String[]> data = new ArrayList<String[]>();
String[] row = \{"222", "Daisy", "3", "334.456", "true", "1956-11-08",
"2013-12-10 12:00:20", "Pune", "IT"};
data.add(row);
writeData("id int,name string,rank short,salary double,active boolean,dob
date,doj timestamp,city string,dept string", data, outputPath, uid, null, null);
String hdfsPath = moveFiles(outputPath, outputPath);
String datapath = hdfsPath.concat("/" + name.getMethodName());
runSQL("create table " + name.getMethodName() + " using carbon location '" +
datapath + "'");
runSQL("select count(*) from " + name.getMethodName());
System.out.println("----Insert------");
runSQL("insert into table " + name.getMethodName() + " select * from " +
name.getMethodName() + "large");
System.out.println("Inserted");
System.out.println("----------After Insert--------------");
System.out.println("----Query 1----");
runSQL("select count(*) from " + name.getMethodName());
// configure cache size = 4 blocklet
CarbonProperties.getInstance()
.addProperty(CarbonCommonConstants.CARBON_MAX_PAGINATION_LRU_CACHE_SIZE_IN_MB,
"4");
CarbonReaderBuilder carbonReaderBuilder = CarbonReader.builder(datapath,
"_temp").withPaginationSupport().projection(new
String[]\{"id","name","rank","salary","active","dob","doj","city","dept"});
PaginationCarbonReader<Object> paginationCarbonReader =
(PaginationCarbonReader<Object>) carbonReaderBuilder.build();
File[] dataFiles1 = new File(datapath).listFiles(new FilenameFilter() {
@Override public boolean accept(File dir, String name) {
return name.endsWith("carbondata");
}
});
String
version=CarbonSchemaReader.getVersionDetails(dataFiles1[0].getAbsolutePath());
System.out.println("version "+version);
System.out.println("Total no of rows is :
"+paginationCarbonReader.getTotalRows() );
assertTrue(paginationCarbonReader.getTotalRows() == 6);
Object[] rows=paginationCarbonReader.read(1,6);
//assertTrue(rows.length==5);
for (Object rowss : rows) {
System.out.println(((Object[]) rowss)[0]);
// assertTrue (((Object[]) row)[1].equals(5001));
}
// close the reader
paginationCarbonReader.close();
}
Issue 2 : when filter () is used to filter certain row . getTotalRows() still
showing previous total no of row where as when try to read all the row getting
error.
/**
* Carbon Files are written using CarbonWriter in outputpath
*
* Carbon Files are read using paginationCarbonReader object with filter
*/
@Test
public void testSDKPaginationFilter() throws IOException,
InvalidLoadOptionException, InterruptedException {
System.out.println("___________________________________________" +
name.getMethodName() + " TestCase Execution is
started________________________________________________");
List<String []> data =new ArrayList<String []>();
String [] row1= \{"100","MNO","A","1001"};
String [] row2= \{"100","MNOP","C","3001"};
String [] row3= \{"100","MNOQ","X","2001"};
String [] row4= \{"100","MNOR","Z","7001"};
String [] row5= \{"100","MNOS","P","5001"};
data.add(row1);
data.add(row2);
data.add(row3);
data.add(row4);
data.add(row5);
String outputPath=getOutputPath(outputDir,name.getMethodName());
boolean isTransactionalTable=false;
long uid=System.currentTimeMillis();
String blockletsize= String.valueOf(2);
String blocksize= String.valueOf(4);
String [] sortColumns=\{"c4","c3"};
writeData("c1 int,c2 string,c3 string,c4 int",data,outputPath,
uid,blocksize,blockletsize,sortColumns);
System.out.println("Data is written");
String hdfsPath = moveFiles(outputPath, outputPath);
String dataPath = hdfsPath.concat("/" + name.getMethodName());
System.out.println("HDFS Data Path is: " + dataPath);
// configure cache size = 4 blocklet
CarbonProperties.getInstance()
.addProperty(CarbonCommonConstants.CARBON_MAX_PAGINATION_LRU_CACHE_SIZE_IN_MB,
"4");
//filter expression
EqualToExpression equalExpression =
new EqualToExpression(new ColumnExpression("c3", DataTypes.STRING),
new LiteralExpression("P", DataTypes.STRING));
CarbonReaderBuilder carbonReaderBuilder = CarbonReader.builder(dataPath,
"_temp").withPaginationSupport().projection(new String[]\{"c2",
"c4"}).filter(equalExpression);
PaginationCarbonReader<Object> paginationCarbonReader =
(PaginationCarbonReader<Object>) carbonReaderBuilder.build();
File[] dataFiles1 = new File(dataPath).listFiles(new FilenameFilter() {
@Override public boolean accept(File dir, String name) {
return name.endsWith("carbondata");
}
});
String
version=CarbonSchemaReader.getVersionDetails(dataFiles1[0].getAbsolutePath());
System.out.println("version "+version);
System.out.println("Total no of rows is :
"+paginationCarbonReader.getTotalRows() );
assertTrue(paginationCarbonReader.getTotalRows() == 5);
Object[] rows=paginationCarbonReader.read(1,5);
for (Object row : rows) {
System.out.println(((Object[]) row)[0]);
// assertTrue (((Object[]) row)[1].equals(5001));
}
// close the reader
paginationCarbonReader.close();
}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
