pyspark-kudu.sh

?

import os
os.environ["JAVA_HOME"] = "/usr/lib/jvm/java-8-openjdk-amd64"
os.environ["SPARK_HOME"] = "/content/spark-2.4.7-bin-hadoop2.7"

#  --driver-class-path {driver_path} --jars {driver_path}

os.environ['PYSPARK_SUBMIT_ARGS'] = f'pyspark-shell --packages org.apache.kudu:kudu-spark2_2.11:1.12.0'

import findspark
findspark.init("spark-2.4.7-bin-hadoop2.7")

from pyspark.sql import SparkSession

spark = SparkSession \
    .builder \
    .appName("Bea Cukai - PySpark/Kudu") \
    .config('spark.driver.bindAddress','localhost') \
    .config('spark.driver.allowMultipleContexts','true') \
    .config('spark.sql.execution.arrow.pyspark.enabled', 'true') \
    .getOrCreate()

table = "transactions_1k"
kudu_master = "178.128.112.105:7051,178.128.112.105:7151,178.128.112.105:7251"
sfmta_kudu = spark.read.option("kudu.master", kudu_master).option("kudu.table", table).option("kudu.scanLocality", "leader_only").format("kudu").load()

sfmta_kudu.createOrReplaceTempView(table)

sdf = spark.sql("SELECT count(*) FROM " + table)

http://178.128.112.105:8050/tablets

pdf = sdf.select("*").toPandas()
pdf.head()

https://www.cioreview.com/hover_img/matiltonbig1.jpg