Pipeline I/O

When you create a pipeline, you'll often need to read data from some external source, such as a file in Google Cloud Storage or a BigQuery table. Likewise, you'll want your pipeline to output its result data to a similar external data sink, such as output files in Cloud Storage or BigQuery. The Dataflow SDKs provide transforms that can read data from an external source or write data to an external sink.

The Dataflow SDKs provide Read and Write transforms for a number of common data storage types. In addition, the Read and Write APIs are extensible; if you want your pipeline to read from or write to a data storage format that isn't supported by the built-in transforms, you can build extensions to provide your own Read and Write operations.

Reading Input Data

Read transforms read data from an external source and return a PCollection representation of the data for use by your pipeline. You can use a Read transform at any point while constructing your pipeline to create a new PCollection, though it will be most common at the start of your program.


Note: Since Read transforms do not have input PCollections, they are applied to the Pipeline directly. As usual, the call to apply returns a PCollection of the appropriate type, whose elements represent the data. See Constructing Your Pipeline for more information.

Reading From Multiple Locations

Many Read transforms, such as Text, support reading from multiple input files matching a glob operator you provide. Consider the following use of the Read transform, which uses a glob operator (*) to read all matching input files in the given location in Google Cloud Storage:



The above Read will read all files at the given location in Cloud Storage with the prefix "input-" and the suffix ".csv".

To read data from disparate sources into a single PCollection, read each one independently and then use the Flatten transform to create a single PCollection.

Writing Output Data

Write transforms write the data in a PCollection to an external data source. You'll most often use Write transforms at the end of your program to output your pipeline's final results. However, you can use Write to output a PCollection's data at any point in your pipeline.

To use a Write transform, you call the apply method on the PCollection that you want to write, and pass the appropriate Write transform as an argument.


When you apply a Write transform to a PCollection, the return value is an object of type PDone. The PDone object is a trivial result object and can be safely ignored.

Writing To Multiple Output Files

For file-based input and output data, such as Text, Write transforms write to multiple output files by default. The Cloud Dataflow service always produces sharded output files automatically. When you pass an output file name to a Write transform, the file name is used as the prefix for all output files that the Write transform produces.

You can append a suffix to each output file by specifying a suffix to your Write transform.

Consider the following use of the Write transform, which writes multiple output files to a location in Cloud Storage. Each file has the prefix "numbers", a numeric tag, and the suffix ".csv".



The above Write will write multiple output files to the given location in Cloud Storage with the prefix "numbers" and the suffix ".csv".

I/O APIs Included in the Dataflow SDKs

Some Source and Sink APIs are included in the Dataflow SDKs.


The Dataflow SDK for Java provides Read and Write transforms for a number of common data formats including:

Additional I/O APIs

In addition to the I/O APIs, the Dataflow SDKs provides an extensible API that you can use to create your own custom data sources and sinks.


You can create your own custom input sources and output sinks using Dataflow's Source and Sink APIs.