import org.apache.spark.sql.DataFrameWriter
val nums: Dataset[Long] = ...
val writer: DataFrameWriter[Row] = nums.write
DataFrameWriter
DataFrameWriter
is an interface to write the result of executing structured query to an external storage system in a batch fashion.
Note
|
Structured Streaming’s DataStreamWriter is responsible for writing in a streaming fashion.
|
DataFrameWriter
is available using write method of a Dataset
.
DataFrameWriter
has a direct support for many file formats, JDBC databases and an extension point to plug in new formats. It assumes parquet as the default data source that you can change using spark.sql.sources.default setting or format method.
// see above for writer definition
// Save dataset in Parquet format
writer.save(path = "nums")
// Save dataset in JSON format
writer.format("json").save(path = "nums-json")
In the end, you trigger the actual saving of the content of a Dataset
using save method.
writer.save
Note
|
DataFrameWriter is really a type constructor in Scala and keeps a reference to a source DataFrame during its lifecycle (starting right from the moment it was created).
|
runCommand
Internal Method
runCommand
(session: SparkSession, name: String)
(command: LogicalPlan): Unit
Caution
|
FIXME |
Internal State
DataFrameWriter
uses the following mutable attributes to build a properly-defined write specification for insertInto, saveAsTable, and save:
Attribute | Setters |
---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
saveAsTable
Method
saveAsTable(tableName: String): Unit
saveAsTable
saves the content of a DataFrame
as the tableName
table.
First, tableName
is parsed to an internal table identifier. saveAsTable
then checks whether the table exists or not and uses save mode to decide what to do.
saveAsTable
uses the SessionCatalog for the current session.
Does table exist? | Save Mode | Behaviour |
---|---|---|
yes |
|
Do nothing |
yes |
|
Throws a |
anything |
anything |
It creates a |
val ints = 0 to 9 toDF
val options = Map("path" -> "/tmp/ints")
ints.write.options(options).saveAsTable("ints")
sql("show tables").show
Saving DataFrame — save
Method
save(): Unit
save
saves the result of a structured query (the content of a Dataset) to a data source.
Internally, save
runs a SaveIntoDataSourceCommand runnable command under the name save
.
Note
|
save does not support saving to Hive (when source is hive ) and bucketing.
|
Caution
|
FIXME What does bucketing mean? What about assertNotBucketed ?
|
Caution
|
FIXME What is partitioningColumns ?
|
Note
|
save uses source , partitioningColumns , extraOptions , and mode internal properties.
|
jdbc
Method
jdbc(url: String, table: String, connectionProperties: Properties): Unit
jdbc
method saves the content of the DataFrame
to an external database table via JDBC.
You can use mode to control save mode, i.e. what happens when an external table exists when save
is executed.
It is assumed that the jdbc
save pipeline is not partitioned and bucketed.
All options are overriden by the input connectionProperties
.
The required options are:
-
driver
which is the class name of the JDBC driver (that is passed to Spark’s ownDriverRegistry.register
and later used toconnect(url, properties)
).
When table
exists and the override save mode is in use, DROP TABLE table
is executed.
It creates the input table
(using CREATE TABLE table (schema)
where schema
is the schema of the DataFrame
).
bucketBy
Method
Caution
|
FIXME |
Defining Write Behaviour Per Sink’s Existence (aka Save Mode) — mode
Method
mode(saveMode: String): DataFrameWriter[T]
mode(saveMode: SaveMode): DataFrameWriter[T]
mode
defines the behaviour of save when an external file or table (Spark writes to) already exists, i.e. SaveMode
.
Name | Description |
---|---|
|
Records are appended to existing data. |
|
Exception is thrown. |
|
Do not save the records and not change the existing data in any way. |
|
Existing data is overwritten by new records. |
Writer Configuration — option
and options
Methods
Caution
|
FIXME |
Writing DataFrames to Files
Caution
|
FIXME |
Specifying Alias or Fully-Qualified Class Name of DataSource — format
Method
Caution
|
FIXME Compare to DataFrameReader. |
insertInto
Method
Caution
|
FIXME |