Hive Commands

Introduction to Hive Commands

Hive command is a data warehouse infrastructure tool that sits on top Hadoop to summarize Big data. It processes structured data. It makes data querying and analyzing easier. Hive command is also called as “schema on reading;” It doesn’t verify data when it is loaded, verification happens only when a query is issued. This property of Hive makes it fast for initial loading. It’s like copying or simply moving a file without putting any constraints or checks. It was first developed by Facebook. Apache Software Foundation took it up later and developed it further.

Components of Hive Commands

The components 0f hive command is as shown below:

Fig 1. Components of Hive

https://www.developer.com/

Features of Hive Commands

Here are few of the features discussed below:

1. Hive stores are raw and processed dataset in Hadoop.
2. It is designed for Online Transaction Processing (OLTP). OLTP is the systems that facilitate high volume data in very less time with no reliance on the single server.
3. It is fast, scalable and reliable.
4. The SQL type querying language provided here is called HiveQL or HQL. This makes ETL tasks and other analysis easier.

Fig 2. Hive Properties

Sources images:- Google

There are few limitations of Hive command as well, which are listed below:

1. Hive doesn’t support subqueries.
2. Hive surely supports over-writing, but unfortunately, it doesn’t support deletion and updates.
3. Hive is not designed for OLTP, but it is used for it.

To enter the Hive’s interactive shell:

$HIVE_HOME/bin/hive

Basic Hive Commands

The basic commands are as explained below.

1. Create

This will create the new database in Hive.

2. Drop

The drop will remove a table from Hive

3. Alter

Alter command will help you rename the table or table columns.

For example:

hive> ALTER TABLE employee RENAME TO employee1;

4. Show

Show command will show all the databases residing in Hive.

5. Describe

Describe command will help you with the information about the schema of the table.

Intermediate Hive Commands

Hive divides a table into variously related partitions based on columns. Using these partitions, it gets easier to query data. These partitions further get divided into buckets, to run query efficiently on to data.

In other words, buckets distribute data into the set of clusters by calculating the hash code of key mentioned in the query.

1. Adding Partition

Adding partition can be accomplished by altering the table. Say you have table “EMP”, with fields such as Id, Name, Salary, Dept, Designation, and yoj.

hive> ALTER TABLE employee

> ADD PARTITION (year=’2012’)

location ‘/2012/part2012’;

2. Renaming Partition

hive> ALTER TABLE employee PARTITION (year=’1203’)

RENAME TO PARTITION (Yoj=’1203’);

3. Drop Partition

hive> ALTER TABLE employee DROP [IF EXISTS]

> PARTITION (year=’1203’);

4. Relational Operators

Relational operators consist of a certain set of operators, that helps in fetching relevant information.

For example: Say your “EMP” table look like this:

Let’s execute Hive query which will fetch us the employee whose salary is greater than 30000.

hive> SELECT * FROM EMP WHERE Salary>=40000;

5. Arithmetic Operators

These are operators which help in help in executing arithmetic operations on the operands, and in turn, always return number types.

For example:  To add two number such as 22 & 33

hive> SELECT 22+33 ADD FROM temp;

6. Logical Operator

These operators are to execute logical operations, which in return always return True/False.

hive> SELECT * FROM EMP WHERE Salary>40000 && Dept=TP;

Advanced Hive Commands

The advanced commands are as explained below.

1. View

View concept in Hive is similar like in SQL. The view can be created at the time of executing a SELECT statement.

Example:

hive> CREATE VIEW EMP_30000 AS

SELECT * FROM EMP

WHERE salary>30000;

2. Loading Data into Table

Hive> Load data local inpath ‘/home/hduser/Desktop/AllStates.csv’ into table States;

Here “States” is the already created table in Hive.

https://www.tutorialspoint.com/hive/

Hive has some built-in functions which help you in fetching your result in a better way.

Like round, floor, BIGINT etc.

4. Join

Join clause can help in joining two tables based on the same column name.

Example:

hive> SELECT c.ID, c.NAME, c.AGE, o.AMOUNT

FROM CUSTOMERS c JOIN ORDERS o

ON (c.ID = o.CUSTOMER_ID);

All kinds of joins are supported by Hive: Left outer join, right outer join, full outer join.

Tips and Tricks to use Hive Commands

Hive makes data processing that easy, straightforward and extensible, that user pay less attention towards optimizing the Hive queries. But paying attention towards few things while writing Hive query, will surely bring great success in managing the workload and saving money. Below are few tips regarding that:

1. Partitions & Buckets: It is a big data tool, which can query on large datasets. However, writing the query without understanding the domain can bring great partitions in Hive.

If the user is aware of the dataset, then relevant and highly used columns could be grouped into the same partition. This will help in running the query faster and inefficient way.

Ultimately the no. of mapper and I/O operations will also be reduced.

Fig 3. Partitioning

Sources images: Google image

Fig 4 Bucketing

Sources images:- Google image

2. Parallel Execution: It runs the query in multiple stages. In some cases these stages may depend on other stages, a hence can’t get started, once the previous stage is completed. However, independent tasks can run parallelly to save overall run time. To enable the parallel run in Hive:

set hive.exec.parallel=true;

Hence, this will enhance the cluster utilization.

3. Block Sampling: Sampling data from a table will allow exploration of queries on data.

Despite bucking, we rather want to sample dataset more randomly. Block sampling comes with various powerful syntax, which helps in sampling the data in various way.

Sampling can be used for finding approx. info from dataset like the average distance between origin and destination.

Querying 1% of big data will give near to the perfect answer. Exploration gets way easier & effective.

Conclusion

Hive is a higher-level abstraction on top of HDFS, which provides flexible query language. It helps in querying and processing data in an easier way.

It can be clubbed with other Big data elements, to harness its functionality in a full-fledged way.

Recommended Articles

This has been a guide to Hive Commands. Here we have discussed components, features, basic, advanced as well as intermediate of Hive Commands along with tips and tricks. You may also look at the following article to learn more –

1. Hive Interview Questions
2. Hive VS Hue – Top 6 Useful Comparisons
3. Tableau Commands
4. Adobe Photoshop Commands
5. Using ORDER BY Function in Hive
6. Download and Install Hive Step by Step
7. Bucketing in Hive | Examples | Features