Skip to contents

Write Parquet file with optional encryption

Source: R/write_parquet.R
write_parquet.Rd

Writes a data frame to a Parquet file. When encryption_key is supplied the file is encrypted with AES using DuckDB's native Parquet encryption support. Without a key the file is written by arrow::write_parquet() and supports any compression codec understood by the Arrow library.

Usage

write_parquet(
  data,
  path,
  ...,
  encryption_key = NULL,
  conn = NULL,
  compression = "zstd"
)

Arguments

data

A data frame or tibble to write.

path

Character string. Destination file path for the Parquet file.

...

Additional arguments passed to arrow::write_parquet() when encryption_key is NULL.

encryption_key

A raw AES key as a hex string (32, 48, or 64 hex characters) or a base-64–encoded 256-bit key string. When NULL (default) no encryption is applied.

conn

An optional existing DuckDB DBIConnection. When supplied, the connection is reused instead of opening a new one, which avoids per-call setup overhead when writing many tables in a batch. The caller is responsible for disconnecting the connection. Default is NULL.

compression

Compression codec to use for unencrypted Parquet files. Any codec supported by arrow::write_parquet() is valid (e.g. "zstd", "snappy", "gzip", "lz4", "uncompressed"). Ignored when encryption_key is set (DuckDB chooses the codec). Default is "zstd".

Value

NULL invisibly. The Parquet file is written to path.

Examples

if (FALSE) { # \dontrun{
data <- mtcars
key <- "rppqM5CuEqotys4wQq/g7xh6wpIjRozcAIbI9sagwKE="

temp_dir <- tempdir()

# Encrypted write
rcdf::write_parquet(
  data = data,
  path = file.path(temp_dir, "mtcars.parquet"),
  encryption_key = key
)

# Unencrypted write with gzip compression
rcdf::write_parquet(
  data = data,
  path = file.path(temp_dir, "mtcars-gz.parquet"),
  compression = "gzip"
)
} # }