Read a TFRecord File

R/loader.R

spark_read_tfrecord

Description

Read a TFRecord file as a Spark DataFrame.

Usage

spark_read_tfrecord(sc, name, path, schema = NULL, 
  record_type = c("Example", "SequenceExample"), overwrite = TRUE) 

Arguments

Arguments Description
sc A spark conneciton.
name The name to assign to the newly generated table.
path The path to the file. Needs to be accessible from the cluster. Supports the “hdfs://”, “s3a://” and “file://” protocols.
schema (Currently unsupported.) Schema of TensorFlow records. If not provided, the schema is inferred from TensorFlow records.
record_type Input format of TensorFlow records. By default it is Example.
overwrite Boolean; overwrite the table with the given name if it already exists?