Each column in a DataFrame has a data type (dtype). Some functions and methods expect columns in a specific data type, and therefore it is a common operation to convert the data type of columns. In this short how-to article, we will learn how to change the data type of a column in Pandas and PySpark DataFrames.
Pandas
In a Pandas DataFrame, we can check the data types of columns with the dtypes method.
df.dtypes
Name string
City string
Age string
dtype: object
The astype function changes the data type of columns. Consider we have a column with numerical values but its data type is string. This is a serious issue because we cannot perform any numerical analysis on textual data.
df["Age"] = df["Age"].astype("int")
We just need to write the desired data type inside the astype function. Let’s confirm the changes by checking the data types again.
df.dtypes
Name string
City string
Age int64
dtype: object
It is possible to change the data type of multiple columns in a single operation. The columns and their data types are written as key-value pairs in a dictionary.
df = df.astype({"Age": "int", "Score": "int"})
PySpark
In PySpark, we can use the cast method to change the data type.
from pyspark.sql.types import IntegerType
from pyspark.sql import functions as F
# first method
df = df.withColumn("Age", df.age.cast("int"))
# second method
df = df.withColumn("Age", df.age.cast(IntegerType()))
# third method
df = df.withColumn("Age", F.col("Age").cast(IntegerType()))
To change the data type of multiple columns, we can combine operations by chaining them.
df = df.withColumn("Age", df.age.cast("int")) \
.withColumn("Score", df.age.cast("int"))
This question is also being asked as:
- How to force all strings to floats?