springml / spark-sftp

Spark connector for SFTP

GitHub

Spark SFTP Connector Library

A library for constructing dataframes by downloading files from SFTP and writing dataframe to a SFTP server

Requirements

This library requires Spark 2.x.

For Spark 1.x support, please check spark1.x branch.

Linking

You can link against this library in your program at the following ways:

Maven Dependency

<dependency>
	<groupId>com.springml</groupId>
	<artifactId>spark-sftp_2.11</artifactId>
	<version>1.1.3</version>
</dependency>

SBT Dependency

libraryDependencies += "com.springml" % "spark-sftp_2.11" % "1.1.3"

Using with Spark shell

This package can be added to Spark using the --packages command line option. For example, to include it when starting the spark shell:

$ bin/spark-shell --packages com.springml:spark-sftp_2.11:1.1.3

Features

This package can be used to construct spark dataframe by downloading the files from SFTP server.

This package can also be used to write spark dataframe as a csv|json|acro tp SFTP server

This library requires following options:

  • path: FTP URL of the file to be used for dataframe construction
  • username: SFTP Server Username.
  • password: (Optional) SFTP Server Password.
  • pem: (Optional) Location of PEM file. Either pem or password has to be specified
  • pemPassphrase: (Optional) Passphrase for PEM file.
  • host: SFTP Host.
  • port: (Optional) Port in which SFTP server is running. Default value 22.
  • fileType: Type of the file. Supported types are csv, txt, json, avro and parquet
  • inferSchema: (Optional) InferSchema from the file content. Currently applicable only for csv fileType
  • header: (Optional) Applicable only for csv fileType. Is the first row in CSV file is header.
  • delimiter: (Optional) Set the field delimiter. Applicable only for csv fileType. Default is comma.
  • quote: (Optional) Set the quote character. Applicable only for csv fileType. Default is ".
  • escape: (Optional) Set the escape character. Applicable only for csv fileType. Default is .
  • multiLine: (Optional) Set the multiline. Applicable only for csv fileType. Default is false.
  • codec: (Optional) Applicable only for csv fileType. Compression codec to use when saving to file. Should be the fully qualified name of a class implementing org.apache.hadoop.io.compress.CompressionCodec or one of case-insensitive shorten names (bzip2, gzip, lz4, and snappy). Defaults to no compression when a codec is not specified.

Scala API

// Construct Spark dataframe using file in FTP server
val df = spark.read.
            format("com.springml.spark.sftp").
            option("host", "SFTP_HOST").
            option("username", "SFTP_USER").
            option("password", "****").
            option("fileType", "csv").
            option("delimiter", ";").
            option("quote", "\"").
            option("escape", "\\").
            option("multiLine", "true").
            option("inferSchema", "true").
            load("/ftp/files/sample.csv")

// Write dataframe as CSV file to FTP server
df.write.
      format("com.springml.spark.sftp").
      option("host", "SFTP_HOST").
      option("username", "SFTP_USER").
      option("password", "****").
      option("fileType", "csv").
      option("delimiter", ";").
      option("codec", "bzip2").
      save("/ftp/files/sample.csv")


// Construct spark dataframe using text file in FTP server
 val df = spark.read.
            format("com.springml.spark.sftp").
            option("host", "SFTP_HOST").
            option("username", "SFTP_USER").
            option("password", "****").
            option("fileType", "txt").
            load("config")
            
 // Construct spark dataframe using xml file in FTP server           
            val df = spark.read.
                 format("com.springml.spark.sftp").
                 option("host", "SFTP_HOST").
                 option("username", "SFTP_USER").
                 option("password", "*****").
                 option("fileType", "xml").
                 option("rowTag", "YEAR").load("myxml.xml")
                 
 // Write dataframe as XML file to FTP server           
           
                 df.write.format("com.springml.spark.sftp").
                 option("host", "SFTP_HOST").
                 option("username", "SFTP_USER").
                 option("password", "*****").
                 option("fileType", "xml").
                 option("rootTag", "YTD").
                 option("rowTag", "YEAR").save("myxmlOut.xml.gz")

Java API

// Construct Spark dataframe using file in FTP server
DataFrame df = spark.read().
					format("com.springml.spark.sftp").
				    option("host", "SFTP_HOST").
				    option("username", "SFTP_USER").
				    option("password", "****").
				    option("fileType", "json").
				    load("/ftp/files/sample.json")

// Write dataframe as CSV file to FTP server
df.write().
      format("com.springml.spark.sftp").
      option("host", "SFTP_HOST").
      option("username", "SFTP_USER").
      option("password", "****").
      option("fileType", "json").
      save("/ftp/files/sample.json");

R API

Spark 1.5+:

if (nchar(Sys.getenv("SPARK_HOME")) < 1) {
  Sys.setenv(SPARK_HOME = "/home/spark")
}
library(SparkR, lib.loc = c(file.path(Sys.getenv("SPARK_HOME"), "R", "lib")))
sparkR.session(master = "local[*]", sparkConfig = list(spark.driver.memory = "2g"))

# Construct Spark dataframe using avro file in FTP server
df <- read.df(path="/ftp/files/sample.avro",
            source="com.springml.spark.sftp",
            host="SFTP_HOST",
            username="SFTP_USER",
            pem="/home/user/mypem.pem",
            fileType="avro")

# Write dataframe as avro file to FTP server
write.df(df,
        path="/ftp/files/sample.avro",
        source="com.springml.spark.sftp",
        host="SFTP_HOST",
        username="SFTP_USER",
        pem="/home/user/mypem.pem",
        fileType="avro")

Note

  1. SFTP files are fetched and written using jsch. It will be executed as a single process
  2. Files from SFTP server will be downloaded to temp location and it will be deleted only during spark shutdown

Building From Source

This library is built with SBT, which is automatically downloaded by the included shell script. To build a JAR file simply run build/sbt package from the project root.