val lines = spark.readStream
.format("socket")
// .schema(StructType(schemas))
.option("host", "10.3.87.23")
.option("port", )
.load()
.selectExpr("CAST(value AS STRING)").as[(String)]DF = lines.map(x => {
val obj = JSON.parseObject(x)
val ls = new util.ArrayList()
Is there a way I can have a JDBC connection open through a streaming job. I
have a foreach which is running once per batch. However, I don’t want to
open the connection for each batch but would rather have a persistent
connection that I can reuse. How can I do this?
Thanks.
Asim
I've done this:
1. foreachPartition
2. Open connection.
3. foreach inside the partition.
4. close the connection.
Slightly crufty, but works. Would love to see a better approach.
Regards,
Ashic.
Date: Fri, 5 Dec 2014 12:32:24 -0500
Subject: Spark Streaming Reusing JDBC Connections
From: asimja