org.apache.parquet.crypto

Interface EncryptionPropertiesFactory

All Known Implementing Classes:

PropertiesDrivenCryptoFactory

public interface EncryptionPropertiesFactory

EncryptionPropertiesFactory interface enables transparent activation of Parquet encryption.

Its customized implementations produce encryption properties for each Parquet file, using the input information available in Parquet file writers: file path, file extended schema - and also Hadoop configuration properties that can pass custom parameters required by a crypto factory. A factory implementation can use or ignore any of these inputs.

The example usage could be as below. 1. Write a class to implement EncryptionPropertiesFactory. 2. Set configuration of "parquet.crypto.factory.class" with the fully qualified name of this class. For example, we can set the configuration in SparkSession as below. SparkSession spark = SparkSession .config("parquet.crypto.factory.class", "xxx.xxx.EncryptionPropertiesClassLoaderImpl")

The implementation of this interface will be instantiated by loadFactory(Configuration).

Field Summary

Fields

Modifier and Type	Field and Description
static String	CRYPTO_FACTORY_CLASS_PROPERTY_NAME
static org.slf4j.Logger	LOG

Method Summary

Modifier and Type	Method and Description
FileEncryptionProperties	getFileEncryptionProperties(org.apache.hadoop.conf.Configuration fileHadoopConfig, org.apache.hadoop.fs.Path tempFilePath, WriteSupport.WriteContext fileWriteContext) Get FileEncryptionProperties object which is created by the implementation of this interface.
static EncryptionPropertiesFactory	loadFactory(org.apache.hadoop.conf.Configuration conf) Load EncryptionPropertiesFactory class specified by CRYPTO_FACTORY_CLASS_PROPERTY_NAME as the path in the configuration

Field Detail

LOG

static final org.slf4j.Logger LOG

CRYPTO_FACTORY_CLASS_PROPERTY_NAME

static final String CRYPTO_FACTORY_CLASS_PROPERTY_NAME

See Also:

Constant Field Values

Method Detail

loadFactory

static EncryptionPropertiesFactory loadFactory(org.apache.hadoop.conf.Configuration conf)

Load EncryptionPropertiesFactory class specified by CRYPTO_FACTORY_CLASS_PROPERTY_NAME as the path in the configuration

Parameters:

conf - Configuration where user specifies the class path

Returns:

object with class EncryptionPropertiesFactory if user specified the class path and invoking of the class succeeds. Null if user doesn't specify the class path (no encryption then).

Throws:

BadConfigurationException - if the instantiation of the configured class fails

getFileEncryptionProperties

FileEncryptionProperties getFileEncryptionProperties(org.apache.hadoop.conf.Configuration fileHadoopConfig,
                                                     org.apache.hadoop.fs.Path tempFilePath,
                                                     WriteSupport.WriteContext fileWriteContext)
                                              throws ParquetCryptoRuntimeException

Get FileEncryptionProperties object which is created by the implementation of this interface. Please see the unit test SampleEncryptionPropertiesFactory for example

Parameters:

fileHadoopConfig - Configuration that is used to pass the needed information, e.g. KMS uri

tempFilePath - File path of the parquet file being written. Can be used for AAD prefix creation, key material management, etc. Implementations must not presume the path is permanent, as the file can be moved or renamed later

fileWriteContext - WriteContext to provide information like schema to build the FileEncryptionProperties

Returns:

object with class of FileEncryptionProperties. Null return value means the file should not be encrypted.

Throws:

ParquetCryptoRuntimeException - if there is an exception while creating the object