How to Use Jdbc Source to Write and Read Data in (Py)Spark

How to use JDBC source to write and read data in (Py)Spark?

Writing data

Include applicable JDBC driver when you submit the application or start shell. You can use for example --packages:
```
 bin/pyspark --packages group:name:version  
```

or combining driver-class-path and jars

    bin/pyspark --driver-class-path $PATH_TO_DRIVER_JAR --jars $PATH_TO_DRIVER_JAR

These properties can be also set using PYSPARK_SUBMIT_ARGS environment variable before JVM instance has been started or using conf/spark-defaults.conf to set spark.jars.packages or spark.jars / spark.driver.extraClassPath.

Choose desired mode. Spark JDBC writer supports following modes:
- append: Append contents of this :class:DataFrame to existing data.
- overwrite: Overwrite existing data.
- ignore: Silently ignore this operation if data already exists.
- error (default case): Throw an exception if data already exists.
Upserts or other fine-grained modifications are not supported
```
 mode = ...
```

Prepare JDBC URI, for example:

 # You can encode credentials in URI or pass
 # separately using properties argument
 # of jdbc method or options

 url = "jdbc:postgresql://localhost/foobar"

(Optional) Create a dictionary of JDBC arguments.
```
 properties = {
     "user": "foo",
     "password": "bar"
 }
```
properties / options can be also used to set supported JDBC connection properties.

Use DataFrame.write.jdbc

 df.write.jdbc(url=url, table="baz", mode=mode, properties=properties)

to save the data (see pyspark.sql.DataFrameWriter for details).

Known issues:

Suitable driver cannot be found when driver has been included using --packages (java.sql.SQLException: No suitable driver found for jdbc: ...)
Assuming there is no driver version mismatch to solve this you can add driver class to the properties. For example:
```
  properties = {
      ...
      "driver": "org.postgresql.Driver"
  }
```
using df.write.format("jdbc").options(...).save() may result in:
java.lang.RuntimeException: org.apache.spark.sql.execution.datasources.jdbc.DefaultSource does not allow create table as select.
Solution unknown.
in Pyspark 1.3 you can try calling Java method directly:
```
  df._jdf.insertIntoJDBC(url, "baz", True)
```

Reading data

Follow steps 1-4 from Writing data

Use sqlContext.read.jdbc:

 sqlContext.read.jdbc(url=url, table="baz", properties=properties)

or sqlContext.read.format("jdbc"):

    (sqlContext.read.format("jdbc")
        .options(url=url, dbtable="baz", **properties)
        .load())

Known issues and gotchas:

Suitable driver cannot be found - see: Writing data
Spark SQL supports predicate pushdown with JDBC sources although not all predicates can pushed down. It also doesn't delegate limits nor aggregations. Possible workaround is to replace dbtable / table argument with a valid subquery. See for example:
- Does spark predicate pushdown work with JDBC?
- More than one hour to execute pyspark.sql.DataFrame.take(4)
- How to use SQL query to define table in dbtable?
By default JDBC data sources loads data sequentially using a single executor thread. To ensure distributed data loading you can:
- Provide partitioning column (must be IntegerType), lowerBound, upperBound, numPartitions.
- Provide a list of mutually exclusive predicates predicates, one for each desired partition.
See:
- Partitioning in spark while reading from RDBMS via JDBC,
- How to optimize partitioning when migrating data from JDBC source?,
- How to improve performance for slow Spark jobs using DataFrame and JDBC connection?
- How to partition Spark RDD when importing Postgres using JDBC?
In a distributed mode (with partitioning column or predicates) each executor operates in its own transaction. If the source database is modified at the same time there is no guarantee that the final view will be consistent.

Where to find suitable drivers:

Maven Repository (to obtain required coordinates for --packages select desired version and copy data from a Gradle tab in a form compile-group:name:version substituting respective fields) or Maven Central Repository:
- PostgreSQL
- MySQL

Other options

Depending on the database specialized source might exist, and be preferred in some cases:

Greenplum - Pivotal Greenplum-Spark Connector
Apache Phoenix - Apache Spark Plugin
Microsoft SQL Server - Spark connector for Azure SQL Databases and SQL Server
Amazon Redshift - Databricks Redshift connector (current versions available only in a proprietary Databricks Runtime. Discontinued open source version, available on GitHub).

PySpark pyspark.sql.DataFrameReader.jdbc() doesn't accept datetime type upperbound argument as document says

It supports.

The issue here is that the spark.read.jdbc method currently only supports parameters upper/lower bounds for integral type columns.

But you can use load method and DataFrameReader.option to specifiy upperBound and lowerBound for other column types date/timestamp :

df = spark.read.format("jdbc") \
    .option("url", "jdbc:mysql://server/db") \
    .option("dbtable", "table_name") \
    .option("user", "user") \
    .option("password", "xxxx") \
    .option("partitionColumn", "EVENT_CAPTURED") \
    .option("lowerBound", "2016-01-01 00:00:00") \
    .option("upperBound", "2016-01-10 00:00:00") \
    .option("numPartitions", "8") \
    .load()

Or by passing dict of options:

df = spark.read.format("jdbc") \
    .options(*sql_conn_params)\
    .load()

You can see all available options and examples here: https://spark.apache.org/docs/latest/sql-data-sources-jdbc.html

PySpark & JDBC: When should I use spark with JDBC?

Yes, you can install spark locally and use JDBC to connect to your databases. Here is an function to help you connect to my-sql, which you can generalize to any JDBC source by changing the JDBC connection string:

def connect_to_sql(
        spark, jdbc_hostname, jdbc_port, database, data_table, username, password
):
    jdbc_url = "jdbc:mysql://{0}:{1}/{2}".format(jdbc_hostname, jdbc_port, database)

    connection_details = {
        "user": username,
        "password": password,
        "driver": "com.mysql.cj.jdbc.Driver",
    }

    df = spark.read.jdbc(url=jdbc_url, table=data_table, properties=connection_details)
    return df

Spark is better at handling big data than Pandas even on local machines but it comes with a performance overhead due to parallelism and distributed computing. It will definitely serve your purpose on the cluster, but local mode should be used for development only.
Rest assured, Spark (installed locally) will push query, limit and transformation limitations and even handle it better if done correctly. Search, Sort, Filter operations are going to be expensive since the DF is a non-indexed distributed data structure.
Unaware of the speed difference between Presto and Spark, have not tried a comparison.

Hope this helps.

Note: Performance improvement is not guaranteed on local machine even with optimal parallel workload. It does not provide opportunity for distribution.

Spark structured streaming from JDBC source

No, there is no such built-in support in Spark Structured Streaming. The main reason is that most of databases doesn't provided an unified interface for obtaining the changes.

It's possible to get changes from some databases using archive logs, write-ahead logs, etc. But it's database-specific. For many databases the popular choice is Debezium that can read such logs and push list of changes into a Kafka, or something similar, from which it could be consumed by Spark.

Loading data from Oracle table using spark JDBC is extremely slow

As your query has lot of filters you don't even need to bring in the whole dataset and then apply filter on it. But you can push this query down to db engine which will in turn filter the data and return back the result for Glue job.

This can be done as explained in https://stackoverflow.com/a/54375010/4326922 and below is an example for mysql which can be applied for oracle too with few changes.

query= "(select ab.id,ab.name,ab.date1,bb.tStartDate from test.test12 ab join test.test34 bb on ab.id=bb.id where ab.date1>'" + args['start_date'] + "') as testresult"

datasource0 = spark.read.format("jdbc").option("url", "jdbc:mysql://host.test.us-east-2.rds.amazonaws.com:3306/test").option("driver", "com.mysql.jdbc.Driver").option("dbtable", query).option("user", "test").option("password", "Password1234").load()

How to Use Jdbc Source to Write and Read Data in (Py)Spark