User-defined functions (UDFs) in Hadoop - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of a simple UDF in Hive

Consider this Hive UDF written in Java. What will be the output when the UDF is applied to the input string "hadoop"?

Hadoop

public class ReverseStringUDF extends org.apache.hadoop.hive.ql.exec.UDF {
    public String evaluate(String input) {
        if (input == null) return null;
        return new StringBuilder(input).reverse().toString();
    }
}

-- Hive query:
SELECT ReverseStringUDF('hadoop');

A 'hadoop'

B 'poohad'

C 'poodah'

D null

💡 Hint

Think about what the StringBuilder.reverse() method does to the input string.

The UDF reverses the input string. 'hadoop' reversed is 'poodah'.

❓ data_output

intermediate

Result of applying a UDF to a Hive table column

Given a Hive table users with a column username, and a UDF ToUpperCaseUDF that converts strings to uppercase, what is the output of this query?

Hadoop

SELECT username, ToUpperCaseUDF(username) AS username_upper FROM users WHERE username = 'alice';

A [('alice', 'alice')]

B []

C [('ALICE', 'ALICE')]

D [('alice', 'ALICE')]

💡 Hint

The UDF changes the case but does not filter rows.

The query selects the row where username is 'alice' and applies the UDF to convert it to uppercase, resulting in 'ALICE'.

🔧 Debug

advanced

Identify the error in this UDF code

This Java UDF code is intended to return the length of a string. What error will it cause when compiled or run?

Hadoop

public class StringLengthUDF extends org.apache.hadoop.hive.ql.exec.UDF {
    public int evaluate(String input) {
        return input.length();
    }
}

A NullPointerException if input is null

B SyntaxError due to missing return type

C ClassNotFoundException at runtime

D No error, runs correctly

💡 Hint

What happens if the input string is null and you call length()?

If the input is null, calling input.length() causes a NullPointerException at runtime.

🧠 Conceptual

advanced

Understanding UDF registration in Hive

Which statement correctly describes how to register a Java UDF in Hive before using it in queries?

A Use <code>ADD JAR 'path_to_jar'; CREATE FUNCTION function_name AS 'class_name'</code>

B Use <code>REGISTER FUNCTION function_name AS 'class_name'</code>

C Use <code>CREATE FUNCTION function_name AS 'class_name' USING JAR 'path_to_jar'</code>

D No registration needed; just call the UDF class directly in queries

💡 Hint

Think about how Hive loads external code and links it to a function name.

In Hive, you first add the JAR containing the UDF with ADD JAR, then create the function with CREATE FUNCTION referencing the class.

🚀 Application

expert

Predict output of a complex UDF with multiple inputs

Given this UDF that concatenates two strings with a dash, what is the output of the Hive query?

Hadoop

public class ConcatDashUDF extends org.apache.hadoop.hive.ql.exec.UDF {
    public String evaluate(String s1, String s2) {
        if (s1 == null) s1 = "";
        if (s2 == null) s2 = "";
        return s1 + "-" + s2;
    }
}

-- Hive query:
SELECT ConcatDashUDF('data', NULL);

A 'data-NULL'

B 'data-'

C '-NULL'

D null

💡 Hint

Check how the UDF handles null inputs before concatenation.

The UDF replaces null inputs with empty strings, so the output is 'data-' when second input is null.