Remove Special Characters from Data using Hive

I have data that has some special characters like carriage returns (\r), newlines (\n), and some non-printable control characters (like ^M). I need to process this data using Hive for some transformations and later load this data to a database like Oracle using Sqoop or Amazon Redshift using PSQL COPY for analysis.

The Sqoop or PSQL COPY fail to load data because of these invalid characters.
I was trying to find how to strip off these characters at Hive before loading to external database.
Hive’s regexp_extract function comes to rescue, but finding the right REGEX pattern was a challenge.
Finally I figured it out after searching on online forums. Good to know is that Hive supports Java REGEX patterns.
My final Hive query is given below.


SELECT regexp_extract(col1, '\\p{Print}*',0)
FROM table1;

Advertisements

Increase Hue Oozie Dashboard Job Entries

On Hue Oozie dashboard, by default you only see 100 entries for workflow jobs.
This won’t help if you have too many jobs running and want to take a look at jobs failed/killed, for example, two days ago.

How to increase the number of entries you see?

Open the file /etc/hue/conf/hue.ini on Hadoop master node.
Update the following entry in the file to a number you wish. Here I increased to 500.

# Maximum of Oozie workflows or coordinators to retrieve in one API call.
oozie_jobs_count=500

And then restart the Hue service using the commands below.

sudo /sbin/stop hue
sudo /sbin/start hue

Note: Those steps for AWS EMR. Other Hadoop distributions might have different ways of setting the config parameters.

HiveServer2 JDBC Client on AWS

Connecting to Hive database can be done using Hive CLI or beeline from a command prompt and programmatically using a JDBC client. The following section describes setting up a JDBC client for HiveServer2 on Eclipse.

Download the required Hive 0.13.1 JDBC jar files from Amazon site in the section titled ‘To download JDBC drivers’. Extract ZIP file and you can see all jar files in the extracted-to folder.

Using Eclipse IDE, create a Java project and configure the Build Path to add external JARs downloaded in the above step.

Use the following program to connect to database on Hive and run the program. It connects to Hive database on Hadoop NameNode and displays the results on the Eclipse console.

import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.ResultSet;
import java.sql.SQLException;
import java.sql.Statement;

public class JdbcClient {
	public static void main(String[] args) {

		try {
			Class.forName("com.amazon.hive.jdbc41.HS2Driver");
		} catch (ClassNotFoundException e) {
			e.printStackTrace();
		}

		Connection conn = null;
		ResultSet rs = null;
		try {
			conn = DriverManager.getConnection("jdbc:hive2://<namenode_ipaddress>:10000/default", "hadoop", "");
			Statement stmt = conn.createStatement();
			String addJarQuery = "ADD JAR /path/on/namenode/to/json-serde-1.3-jar-with-dependencies.jar";
			String queryStr = "select * from my_table limit 10";

			stmt.execute(addJarQuery);
			rs = stmt.executeQuery(queryStr);

			while (rs.next()) {
				System.out.println(rs.getString(1));
			}
		} catch (SQLException e) {
			e.printStackTrace();
		} finally {
			try {
				if (conn != null)
					conn.close();

			} catch (SQLException e) {
				e.printStackTrace();
			}
		}

		System.out.println("End of the program");
	}
}

P.S: In the above code, if ‘ADD JAR’ Hive command is not there, you get the following error.

java.lang.ClassNotFoundException Class org.openx.data.jsonserde.JsonSerDe

Drop Hive Partitions with Special Characters

I have a partition for a Hive table created wrongly. See the query below. Hive uses URL-encoded values for special characters.

$ hive -e 'show partitions mytable'
date=$%7Bhiveconf%3AINPUT_DATE}

How to drop it?

It will not work if you use the same value displayed above to drop it, even if Hive says OK.

hive> alter table mytable drop partition (date='$%7Bhiveconf%3AINPUT_DATE}');
OK
Time taken: 0.804 seconds

You know the actual partition when you created. If don’t remember check here to know what is the equivalent value for each encoded character value, and use the actual value to drop it.

hive> alter table mytable drop partition (date='${hiveconf:INPUT_DATE}');

 

How to stop Tez jobs

Apache Tez is a fast data processing engine that can be used as an alternative to slow and old MapReduce. Tez runs on YARN.

MapReduce applications (jobs) can be listed using command: mapred job -list

MapReduce jobs can be killed using: mapred job -kill <JobId>

$ mapred job -list
16/03/16 17:22:41 INFO client.RMProxy: Connecting to ResourceManager at ip-10-10-10-10.ec2.internal/10.10.10.10:8032
Total jobs:2
 JobId State StartTime UserName Queue Priority UsedContainers RsvdContainers UsedMem RsvdMem NeededMem AM info
job_1457636583888_30685 RUNNING 1458148947326 hadoop default NORMAL 1 0 3392M 0M 3392M http://ip-10-10-10-10.ec2.internal:20888/proxy/application_1457636583888_30685/
job_1457636583888_30686 PREP 1458148961271 hadoop default NORMAL 1 0 2880M 0M 2880M http://ip-10-10-10-10.ec2.internal:20888/proxy/application_1457636583888_30686/

 

Tez jobs can be listed using: yarn application -list

Tez jobs can be killed using: yarn application -kill <Application-Id>

$ yarn application -list
16/03/16 17:24:25 INFO client.RMProxy: Connecting to ResourceManager at ip-10-10-10-10.ec2.internal/10.10.10.10:8032
Total number of applications (application-types: [] and states: [SUBMITTED, ACCEPTED, RUNNING]):2
 Application-Id Application-Name Application-Type User Queue State Final-State Progress Tracking-URL
application_1457636583888_30638 HIVE-64481d5d-eb4f-4a13-a3d8-ab2584547368 TEZ hadoop default RUNNING UNDEFINED 99.85% http://ip-10-10-10-10.ec2.internal:52648/ui/
application_1457636583888_30690 oozie:launcher:T=hive:W=some_job:A=some_name:ID=0005123-160310193446393-oozie-oozi-W MAPREDUCE hadoop default RUNNING UNDEFINED 5% http://ip-10-10-10-10.ec2.internal:32777

 

 

What is training and model in Machine Learning

Machine Learning is about predicting future behavior based on past data. And this is done by machines (computers), hence the name Machine Learning. It’s also called Data Science in recent times.

What is model?

A model is an algorithm designed to draw some conclusions based on past data.

Example

Based on historical data of payments, it can be predicted the same behavior on a new person.

A person paying monthly rent, credit card, and mortgage payments on time => Another person paying rent, credit payment on time can pay mortgage payments on time too.

What is training a model?

Training a model is designing (computing) an algorithm based on some training data (sample data used to train).

 

P.S.: There is lot to learn on Machine Learning. These two words always confused me. It took long time to understand them.

 

Sqoop Export a Long Text Field

I have a text field on HDFS that can have a very long value of more than 4000 characters length.

I had to export this to Oracle using Sqoop. In Oracle table, this field is defined with data type VARCHAR2(4000 BYTE).

I get an error when Sqoop’ing.

Caused by: java.sql.SQLException: ORA-01461: can bind a LONG value only for insert into a LONG column

Why?

Oracle treats value longer than the defined limit for that field as LONG. Hence the error. The error is not very informative though.

Solution:

Use a CLOB datatype for that field on Oracle. CLOB can store value longer than 4000 characters.

Don’t forget to add –map-column-java option to Sqoop export command. As there is no COLB type in Java/Hive. Hive has to know how to treat this value. Full command is shown below.

sqoop export --connect jdbc:oracle:thin:@hostname:1521/sid \
   --username user --password passwd \ 
   --table Oracle_Table_Name \
   --columns col1,col2,col3 \
   --map-column-java col3=String \
   --export-dir 'location' \
   --input-fields-terminated-by '\001' --input-lines-terminated-by "\n" \
   --input-null-string '\\N' --input-null-non-string '\\N'