I think what you want to achieve is what PySpark is actually doing in
it's API under the hood.
So, specifically you need to look at PySpark's implementation of
DataFrame, SparkSession and SparkContext API. Under the hood that what
is happening, it start a py4j gateway and delegates all Spark operations
and objects creation to JVM.
Look for example here
<https://github.com/apache/spark/blob/48ef9bd2b3a0cb52aaa31a3fada8779b7a7b9132/python/pyspark/context.py#L209>,
here
<https://github.com/apache/spark/blob/48ef9bd2b3a0cb52aaa31a3fada8779b7a7b9132/python/pyspark/sql/session.py#L248>
and here
<https://github.com/apache/spark/blob/48ef9bd2b3a0cb52aaa31a3fada8779b7a7b9132/python/pyspark/java_gateway.py#L153>where
SparkSession/SparkContext (Python) communicates with JVM and creates
SparkSession/SparkContext on JVM side. And rest of the PySpark code will
be delegating execution to them.
Having these objects created by you and custom Java/Scala application
which holds Spark objects then you can play around with rest of
DataFrame creation and passing back and forward. But, I must admit, this
is not part of official documentation and playing around internals of
Spark and which are subject to change (often).
So, I am not sure what your actual requirement is but you will need to
implement your custom version of PySpark API to get all functionality
you need and control on JVM side.
On 31/03/2021 06:49, Aditya Singh wrote:
Thanks a lot Khalid for replying.
I have one question though. The approach tou showed needs an
understanding on python side before hand about the data type of
columns of dataframe. Can we implement a generic approach where this
info is not required and we just have the java dataframe as input on
python side?
Also one more question, in my use case I will sending a dataframe from
java to python and then on python side there will be some
transformation done on the dataframe(including using python udfs) but
no actions will be performed here and then will send it back to java
where actions will be performed. So also wanted to ask if this is
feasible and if yes do we need to send some special jars to executors
so that it can execute udfs on the dataframe.
On Wed, 31 Mar 2021 at 3:37 AM, Khalid Mammadov
<khalidmammad...@gmail.com <mailto:khalidmammad...@gmail.com>> wrote:
Hi Aditya,
I think you original question was as how to convert a DataFrame
from Spark session created on Java/Scala to a DataFrame on a Spark
session created from Python(PySpark).
So, as I have answered on your SO question:
There is a missing call to *entry_point* before calling getDf() in
your code
So, try this :
|app = gateway.entry_point j_df = app.getDf() |
Additionally, I have create working copy using Python and Scala
(hope you dont mind) below that shows how on Scala side py4j
gateway is started with Spark session and a sample DataFrame and
on Python side I have accessed that DataFrame and converted to
Python List[Tuple] before converting back to a DataFrame for a
Spark session on Python side:
*Python:*
|from py4j.java_gateway import JavaGateway from pyspark.sql import
SparkSession from pyspark.sql.types import StructType,
IntegerType, StructField if __name__ == '__main__': gateway =
JavaGateway() spark_app = gateway.entry_point df = spark_app.df()
# Note "apply" method here comes from Scala's companion object to
access elements of an array df_to_list_tuple = [(int(i.apply(0)),
int(i.apply(1))) for i in df] spark = (SparkSession .builder
.appName("My PySpark App") .getOrCreate()) schema = StructType([
StructField("a", IntegerType(), True), StructField("b",
IntegerType(), True)]) df =
spark.createDataFrame(df_to_list_tuple, schema) df.show() |
*Scala:*
|import java.nio.file.{Path, Paths} import
org.apache.spark.sql.SparkSession import py4j.GatewayServer object
SparkApp { val myFile: Path =
Paths.get(System.getProperty("user.home") +
"/dev/sample_data/games.csv") val spark = SparkSession.builder()
.master("local[*]") .appName("My app") .getOrCreate() val df =
spark .read .option("header", "True") .csv(myFile.toString)
.collect() } object Py4JServerApp extends App { val server = new
GatewayServer(SparkApp) server.start() print("Started and
running...") } |
Regards,
Khalid
On 30/03/2021 07:57, Aditya Singh wrote:
Hi Sean,
Thanks a lot for replying and apologies for the late reply(I
somehow missed this mail before) but I am under the impression
that passing the py4j.java_gateway.JavaGateway object lets the
pyspark access the spark context created on the java side.
My use case is exactly what you mentioned in the last email. I
want to access the same spark session across java and pyspark. So
how can we share the spark context and in turn spark session,
across java and pyspark.
Regards,
Aditya
On Fri, 26 Mar 2021 at 6:49 PM, Sean Owen <sro...@gmail.com
<mailto:sro...@gmail.com>> wrote:
The problem is that both of these are not sharing a
SparkContext as far as I can see, so there is no way to share
the object across them, let alone languages.
You can of course write the data from Java, read it from Python.
In some hosted Spark products, you can access the same
session from two languages and register the DataFrame as a
temp view in Java, then access it in Pyspark.
On Fri, Mar 26, 2021 at 8:14 AM Aditya Singh
<aditya.singh9...@gmail.com
<mailto:aditya.singh9...@gmail.com>> wrote:
Hi All,
I am a newbie to spark and trying to pass a java
dataframe to pyspark. Foloowing link has details about
what I am trying to do:-
https://stackoverflow.com/questions/66797382/creating-pysparks-spark-context-py4j-java-gateway-object
<https://stackoverflow.com/questions/66797382/creating-pysparks-spark-context-py4j-java-gateway-object>
Can someone please help me with this?
Thanks,
