How to add any new library like spark-csv in Apache Spark prebuilt version

38.2k views Asked by At

I have build the Spark-csv and able to use the same from pyspark shell using the following command

bin/spark-shell --packages com.databricks:spark-csv_2.10:1.0.3

error getting

>>> df_cat.save("k.csv","com.databricks.spark.csv")
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/abhishekchoudhary/bigdata/cdh5.2.0/spark-1.3.1/python/pyspark/sql/dataframe.py", line 209, in save
    self._jdf.save(source, jmode, joptions)
  File "/Users/abhishekchoudhary/bigdata/cdh5.2.0/spark-1.3.1/python/lib/py4j-0.8.2.1-src.zip/py4j/java_gateway.py", line 538, in __call__
  File "/Users/abhishekchoudhary/bigdata/cdh5.2.0/spark-1.3.1/python/lib/py4j-0.8.2.1-src.zip/py4j/protocol.py", line 300, in get_return_value
py4j.protocol.Py4JJavaError

Where should I place the jar file in my spark pre-built setup so that I will be able to access spark-csv from python editor directly as well.

6

There are 6 answers

1
Yannick Marcon On BEST ANSWER

At the time I used spark-csv, I also had to download commons-csv jar (not sure it is still relevant). Both jars where in the spark distribution folder.

  1. I downloaded the jars as follow:

    wget http://search.maven.org/remotecontent?filepath=org/apache/commons/commons-csv/1.1/commons-csv-1.1.jar -O commons-csv-1.1.jar<br/>    
    wget http://search.maven.org/remotecontent?filepath=com/databricks/spark-csv_2.10/1.0.0/spark-csv_2.10-1.0.0.jar -O spark-csv_2.10-1.0.0.jar
    
  2. then started the python spark shell with the arguments:

    ./bin/pyspark --jars "spark-csv_2.10-1.0.0.jar,commons-csv-1.1.jar"
    
  3. and read a spark dataframe from a csv file:

    from pyspark.sql import SQLContext
    sqlContext = SQLContext(sc)
    df = sqlContext.load(source="com.databricks.spark.csv", path = "/path/to/you/file.csv")
    df.show()
    
0
kentt On

Another option is to add the following to your spark-defaults.conf:

spark.jars.packages com.databricks:spark-csv_2.11:1.2.0
0
Shubham Sinha On

first find out the path of the spark. for example for pyspark

    which pyspark

it will return you the path for example like this- /home/ubuntu/bin/pyspark

then run this command by change the path as per your spark path general-: path --packages com.databricks:spark-csv_2.10:1.0.3

    /home/ubuntu/bin/pyspark --packages com.databricks:spark-csv_2.10:1.0.3
2
Kamil Sindi On

Assuming the session/context hasn't been created yet:

import os

os.environ['PYSPARK_SUBMIT_ARGS'] = '--packages com.databricks:spark-csv_2.10:1.3.0 pyspark-shell'
6
Jimmy On

Instead of placing the jars in any specific folder a simple fix would be to start the pyspark shell with the following arguments:

bin/pyspark --packages com.databricks:spark-csv_2.10:1.0.3

This will automatically load the required spark-csv jars.

Then do the following to read the csv file:

from pyspark.sql import SQLContext
sqlContext = SQLContext(sc)
df = sqlContext.read.format('com.databricks.spark.csv').options(header='true').load('file.csv')
df.show()
0
user3815062 On

Below command helped me -: With Scala 2.10 version

/opt/mapr/spark/spark-1.5.2/bin/spark-shell --master local[*] --packages com.databricks:spark-csv_2.10:1.4.0

Has below dependencies -:

com.databricks#spark-csv_2.10;1.4.0!spark-csv_2.10.jar (2043ms)
org.apache.commons#commons-csv;1.1!commons-csv.jar (419ms)
com.univocity#univocity-parsers;1.5.1!univocity-parsers.jar (1481ms)