Input/output#

CSV#

`read_csv`(filepath_or_buffer[, sep, ...])	Load a comma-separated-values (CSV) dataset into a DataFrame
`DataFrame.to_csv`([path_or_buf, sep, na_rep, ...])	Write a dataframe to csv file format.

read_text(filepath_or_buffer[, delimiter, ...])

Configuration object for a text Datasource

`read_json`(path_or_buf[, engine, orient, ...])	Load a JSON dataset into a DataFrame
`DataFrame.to_json`([path_or_buf])	Convert the cuDF object to a JSON string.

`read_parquet`(filepath_or_buffer[, engine, ...])	Load a Parquet dataset into a DataFrame
`DataFrame.to_parquet`(path[, engine, ...])	Write a DataFrame to the parquet format.
`cudf.io.parquet.read_parquet_metadata`(path)	Read a Parquet file's metadata and schema
`cudf.io.parquet.ParquetDatasetWriter`(path, ...)	Write a parquet file or dataset incrementally

`read_orc`(filepath_or_buffer[, engine, ...])	Load an ORC dataset into a DataFrame
`DataFrame.to_orc`(fname[, compression, ...])	Write a DataFrame to the ORC format.

`read_hdf`(path_or_buf, args, *kwargs)	Read from the store, close it if we opened it.
`DataFrame.to_hdf`(path_or_buf, key, *args, ...)	Write the contained data to an HDF5 file using HDFStore.

Warning

HDF reader and writers are not GPU accelerated. These currently use CPU via Pandas. This may be GPU accelerated in the future.

`read_feather`(path, args, *kwargs)	Load an feather object from the file path, returning a DataFrame.
`DataFrame.to_feather`(path, args, *kwargs)	Write a DataFrame to the feather format.

Warning

Feather reader and writers are not GPU accelerated. These currently use CPU via Pandas. This may be GPU accelerated in the future.

read_avro(filepath_or_buffer[, columns, ...])

Load an Avro dataset into a DataFrame