String functions in Spark in Apache Spark - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of substring extraction in Spark

What is the output of this Spark code snippet that extracts a substring from a column?

Apache Spark

from pyspark.sql import SparkSession
from pyspark.sql.functions import substring

spark = SparkSession.builder.getOrCreate()
data = [(1, "DataScience"), (2, "SparkFun"), (3, "Python")]
df = spark.createDataFrame(data, ["id", "word"])
df.select("id", substring("word", 2, 4).alias("sub_word")).orderBy("id").show()

A

+---+--------+
|id |sub_word|
+---+--------+
|1  |ataS    |
|2  |park    |
|3  |yth     |
+---+--------+

B

+---+--------+
|id |sub_word|
+---+--------+
|1  |ataS    |
|2  |par     |
|3  |ytho    |
+---+--------+

C

+---+--------+
|id |sub_word|
+---+--------+
|1  |ataS    |
|2  |parkF   |
|3  |ytho    |
+---+--------+

D

+---+--------+
|id |sub_word|
+---+--------+
|1  |ataS    |
|2  |park    |
|3  |ytho    |
+---+--------+

💡 Hint

Remember that substring in Spark starts at position 1 and length is the number of characters to extract.

The substring function extracts 4 characters starting from position 2 (1-based index). For 'DataScience', characters 2 to 5 are 'ataS'. For 'SparkFun', characters 2 to 5 are 'park'. For 'Python', characters 2 to 5 are 'ytho'.

❓ data_output

intermediate

Result of trimming spaces in Spark DataFrame

Given this Spark DataFrame, what is the output after applying the trim function to the 'text' column?

Apache Spark

from pyspark.sql import SparkSession
from pyspark.sql.functions import trim

spark = SparkSession.builder.getOrCreate()
data = [(1, "  hello  "), (2, " world"), (3, "spark  ")]
df = spark.createDataFrame(data, ["id", "text"])
df.select("id", trim("text").alias("trimmed_text")).orderBy("id").show()

A

+---+------------+
|id |trimmed_text|
+---+------------+
|1  |hello       |
|2  |world       |
|3  |spark       |
+---+------------+

B

+---+------------+
|id |trimmed_text|
+---+------------+
|1  |  hello     |
|2  | world      |
|3  |spark       |
+---+------------+

C

+---+------------+
|id |trimmed_text|
+---+------------+
|1  |hello       |
|2  |world       |
|3  |spark  	   |
+---+------------+

D

+---+------------+
|id |trimmed_text|
+---+------------+
|1  |hello       |
|2  |world       |
|3  | spark      |
+---+------------+

💡 Hint

The trim function removes spaces from both ends of the string.

The trim function removes leading and trailing spaces. So ' hello ' becomes 'hello', ' world' becomes 'world', and 'spark ' becomes 'spark'.

🔧 Debug

advanced

Identify the error in Spark string concatenation

What error will this Spark code raise when trying to concatenate two string columns?

Apache Spark

from pyspark.sql import SparkSession
from pyspark.sql.functions import concat

spark = SparkSession.builder.getOrCreate()
data = [(1, "Hello", "World"), (2, "Spark", "Fun")]
df = spark.createDataFrame(data, ["id", "col1", "col2"])
df.select("id", concat("col1", "col2").alias("greeting")).show()

A No error, outputs concatenated strings without space

B AnalysisException: cannot resolve 'concat(col1, col2)' due to data type mismatch

C TypeError: concat() argument must be a Column or a list of Columns

D SyntaxError: invalid syntax in concat function call

💡 Hint

Check the argument types passed to concat function in Spark.

The concat function expects Column objects, not string column names. Passing strings causes a TypeError.

🚀 Application

advanced

Using regexp_replace to clean data

You want to remove all digits from the 'info' column in a Spark DataFrame. Which code snippet correctly does this?

A df.withColumn('clean_info', regexp_replace('info', '\\d', '')).show()

B df.withColumn('clean_info', regexp_replace('info', '[0-9]+', '')).show()

C df.withColumn('clean_info', regexp_replace('info', '[a-z]', '')).show()

D df.withColumn('clean_info', regexp_replace('info', '\\D', '')).show()

💡 Hint

Digits are represented by \d or [0-9] in regex, but consider the difference between \d and \D.

Option B uses '[0-9]+' which matches one or more digits and replaces them with empty string, effectively removing digits. Option B uses '\d' but the backslash needs to be escaped properly in Python strings. Option B removes letters, not digits. Option B removes non-digits, which is the opposite of what is wanted.

🧠 Conceptual

expert

Understanding Spark string function behavior with nulls

What is the result of applying the Spark function upper() to a column containing null values?

A The upper() function returns null for null input values.

B The upper() function raises a NullPointerException when encountering null values.

C The upper() function converts null values to empty strings before converting to uppercase.

D The upper() function ignores null values and leaves them unchanged in the output.

💡 Hint

Think about how Spark functions handle null inputs generally.

Spark string functions like upper() return null if the input is null. They do not raise errors or convert nulls to empty strings.