Spark dataframe save in single file on hdfs location

shikha dubey picture shikha dubey · Nov 24, 2016 · Viewed 15.9k times · Source

I have dataframe and i want to save in single file on hdfs location.

i found the solution here Write single CSV file using spark-csv

df.coalesce(1)
    .write.format("com.databricks.spark.csv")
    .option("header", "true")
    .save("mydata.csv")

But all data will be written to mydata.csv/part-00000 and i wanted to be mydata.csv file.

is that possible?

any help appreciate

Answer

Mariusz picture Mariusz · Nov 24, 2016

It's not possible using standard spark library, but you can use Hadoop API for managing filesystem - save output in temporary directory and then move file to the requested path. For example (in pyspark):

df.coalesce(1) \
    .write.format("com.databricks.spark.csv") \
    .option("header", "true") \
    .save("mydata.csv-temp")

from py4j.java_gateway import java_import
java_import(spark._jvm, 'org.apache.hadoop.fs.Path')

fs = spark._jvm.org.apache.hadoop.fs.FileSystem.get(spark._jsc.hadoopConfiguration())
file = fs.globStatus(sc._jvm.Path('mydata.csv-temp/part*'))[0].getPath().getName()
fs.rename(sc._jvm.Path('mydata.csv-temp/' + file), sc._jvm.Path('mydata.csv'))
fs.delete(sc._jvm.Path('mydata.csv-temp'), True)