Null and duplicate detection in Apache Spark - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Detecting Null Values Count in Spark DataFrame

What is the output of this Spark code snippet that counts null values in each column?

Apache Spark

from pyspark.sql import SparkSession
from pyspark.sql.functions import col, sum

spark = SparkSession.builder.getOrCreate()
data = [(1, None), (2, 'a'), (None, 'b'), (4, None)]
df = spark.createDataFrame(data, ['id', 'value'])

null_counts = df.select([sum(col(c).isNull().cast('int')).alias(c) for c in df.columns])
null_counts.show()

A

+---+-----+
| id|value|
+---+-----+
|  1|    2|
+---+-----+

B

+---+-----+
| id|value|
+---+-----+
|  2|    1|
+---+-----+

C

+---+-----+
| id|value|
+---+-----+
|  1|    1|
+---+-----+

D

+---+-----+
| id|value|
+---+-----+
|  0|    2|
+---+-----+

💡 Hint

Remember that isNull() returns a boolean, which we cast to int to sum nulls.

The code counts nulls in each column by checking isNull(), casting True to 1 and False to 0, then summing. The 'id' column has one null, 'value' has two nulls.

❓ data_output

intermediate

Number of Duplicate Rows in Spark DataFrame

Given this Spark DataFrame, how many duplicate rows does it contain?

Apache Spark

from pyspark.sql import SparkSession

spark = SparkSession.builder.getOrCreate()
data = [(1, 'a'), (2, 'b'), (1, 'a'), (3, 'c'), (2, 'b'), (4, 'd')]
df = spark.createDataFrame(data, ['id', 'value'])
duplicates_count = df.count() - df.dropDuplicates().count()
print(duplicates_count)

A 2

B 3

C 1

D 4

💡 Hint

Duplicates are rows that appear more than once.

The DataFrame has 6 rows. After dropping duplicates, 4 unique rows remain. So duplicates count is 6 - 4 = 2.

🔧 Debug

advanced

Identify the Error in Null Filtering Code

What error does this Spark code raise when trying to filter rows with nulls in 'age' column?

Apache Spark

from pyspark.sql import SparkSession

spark = SparkSession.builder.getOrCreate()
data = [(1, 25), (2, None), (3, 30)]
df = spark.createDataFrame(data, ['id', 'age'])

filtered = df.filter(df['age'] == None)
filtered.show()

A No error, but filtered DataFrame is empty

B SyntaxError

C TypeError

D The filter returns no rows because '==' does not detect nulls

💡 Hint

In Spark, '==' does not work to check null values.

Using '==' to compare with None returns no rows because null comparisons require 'isNull()' method in Spark.

🚀 Application

advanced

Remove Duplicate Rows Based on Subset of Columns

Which code snippet correctly removes duplicates based only on the 'email' column in a Spark DataFrame?

Apache Spark

from pyspark.sql import SparkSession

spark = SparkSession.builder.getOrCreate()
data = [(1, '[e-mail]'), (2, '[e-mail]'), (3, '[e-mail]'), (4, '[e-mail]')]
df = spark.createDataFrame(data, ['id', 'email'])

A df.dropDuplicates(['id']).show()

B df.dropDuplicates(['email']).show()

C df.dropDuplicates().show()

D df.dropDuplicates('email').show()

💡 Hint

dropDuplicates expects a list of column names to consider for duplicates.

Option B correctly passes a list with 'email' to dropDuplicates, removing rows with duplicate emails.

🧠 Conceptual

expert

Understanding Null Handling in Spark Aggregations

When using Spark's aggregation functions like sum() on a column with null values, what is the behavior?

A Null values are treated as zero in the aggregation.

B Null values cause the aggregation to return null.

C Null values are ignored in the aggregation, sum returns the total of non-null values.

D Aggregation raises an error if null values exist.

💡 Hint

Think about how SQL aggregates handle nulls.

Spark aggregation functions ignore nulls by default, so sum() adds only non-null values.