Save DataFrame as CSV File in Spark

Spark provides rich APIs to save data frames to many different formats of files such as CSV, Parquet, Orc, Avro, etc. CSV is commonly used in data application though nowadays binary formats are getting momentum. In this article, I am going to show you how to save Spark data frame as CSV file in both local file system and HDFS.

Spark CSV parameters

Refer to the following official documentation about all the parameters supported by CSV api in PySpark.

https://spark.apache.org/docs/latest/api/python/pyspark.sql.html?highlight=savemode#pyspark.sql.DataFrameReader.csv

Example code

In the following sample code, a data frame is created from a python list.  The data frame is then saved to both local file path and HDFS. To save file to local path, specify 'file://'. By default, the path is HDFS path. There are also several options used:

1. header: to specify whether include header in the file.
2. sep: to specify the delimiter
3. mode is used to specify the behavior of the save operation when data already exists.

   • append: Append contents of this DataFrame to existing data.

   • overwrite: Overwrite existing data.

   • ignore: Silently ignore this operation if data already exists.

   • error or errorifexists (default case): Throw an exception if data already exists.

from pyspark.sql import SparkSession
from pyspark.sql.types import ArrayType, StructField, StructType, StringType, IntegerType, DecimalType
from decimal import Decimal

appName = "Python Example - PySpark Save DataFrame as CSV"
master = 'local'

# Create Spark session
spark = SparkSession.builder \
    .master(master) \
    .appName(appName) \
    .getOrCreate()

# List
data = [('Category A', 1, Decimal(12.40)),
        ('Category B', 2, Decimal(30.10)),
        ('Category C', 3, Decimal(100.01)),
        ('Category A', 4, Decimal(110.01)),
        ('Category B', 5, Decimal(70.85))
        ]

# Create a schema for the dataframe
schema = StructType([
    StructField('Category', StringType(), False),
    StructField('ItemID', IntegerType(), False),
    StructField('Amount', DecimalType(scale=2), True)
])

# Convert list to data frame
df = spark.createDataFrame(data, schema)
df.show()

# Save file local folder, delimiter by default is ,
df.write.format('csv').option('header',True).mode('overwrite').option('sep',',').save('file:///home/tangr/output.csv')

# Save file to HDFS
df.write.format('csv').option('header',True).mode('overwrite').option('sep','|').save('/output.csv')

Check the results

You can then check the results in HDFS and local file storage.

The following are examples from my WSL:

tangr@raymond-pc:~$ hadoop fs -ls /
Found 4 items
drwxr-xr-x   - tangr supergroup          0 2019-12-03 20:40 /output.csv
drwxr-xr-x   - tangr supergroup          0 2019-08-25 12:11 /scripts
drwxrwxr-x   - tangr supergroup          0 2019-05-18 15:52 /tmp
drwxr-xr-x   - tangr supergroup          0 2019-08-25 09:35 /user
tangr@raymond-pc:~$ hadoop fs -ls /output.csv
Found 2 items
-rw-r--r--   1 tangr supergroup          0 2019-12-03 20:40 /output.csv/_SUCCESS
-rw-r--r--   1 tangr supergroup        120 2019-12-03 20:40 /output.csv/part-00000-508be2a7-a564-4603-b77c-f4de7c07dbcd-c000.csv
tangr@raymond-pc:~$ hadoop fs -cat /output.csv/part-00000-508be2a7-a564-4603-b77c-f4de7c07dbcd-c000.csv
Category|ItemID|Amount
Category A|1|12.40
Category B|2|30.10
Category C|3|100.01
Category A|4|110.01
Category B|5|70.85
tangr@raymond-pc:~$ cd output.csv/
tangr@raymond-pc:~/output.csv$ ls
_SUCCESS  part-00000-bfbb44b0-1880-4400-a9c1-9c03180553a2-c000.csv
tangr@raymond-pc:~/output.csv$ cat part-00000-bfbb44b0-1880-4400-a9c1-9c03180553a2-c000.csv
Category,ItemID,Amount
Category A,1,12.40
Category B,2,30.10
Category C,3,100.01
Category A,4,110.01
Category B,5,70.85