• Well-known, popular data analytics library
• Many read and write handlers to import/export data
• IO Tools: https://pandas.pydata.org/pandas-docs/stable/user_guide/io.html

• Binary, compressed formats (.bzip2, .gz, .xz) for any R object(s)
• RData/rda types:
  • save()/load()
• rds types
  • saveRDS()/readRDS()

# MULTIPLE OBJECTS
save(df1, df2, df3, file="/path/to/myrdata.rda", compress="gzip")
load("/path/to/myrdata.rda")

# INDIVIDUAL OBJECTS
saveRDS(df1, file="/path/to/df1.rds", compression="bzip2")
new_df1 <- readRDS("/path/to/df1.rds")

saveRDS(df2, file="/path/to/df2.rds", compression="xz")
new_df2 <- readRDS("/path/to/df2.rds")

saveRDS(df3, file="/path/to/df3.rds", compression=FALSE)
new_df3 <- readRDS("/path/to/df3.rds")

• Lightweight C library: librdata
• Read/write data frames stored R data types
• Python + C => Cython => compiled .so/.pyd => import rdata

• Cython extension (compiled in Linux, Windows, Mac)
• Handles compression, http/ftp/s3 paths, file-like objects
• Signature of module, pandas.io.rdata:

import pandas as pd

# READ RDATA
dfs = pd.read_rdata("/path/to/rds_file.rds")

dfs = pd.read_rdata("/path/to/rda_file.rda", 
  file_format="rda", rownames = False,
  select_frames=["df1", "df3", "df5"],
)
# WRITE RDATA
df1.to_rdata("/path/to/new/rds_file.rds")
df2.to_rdata("/path/to/new/rda_file.rda", 
    rda_name="my_df", index=False, compression="bz2"
)

• Documentation: alternative user_guide.rst
• Unit tests: prepared test data prep_test_rdata.R

• Continuous integration across all OSes and Python versions
• Code Standards: C files linting, mypy, pytests
• Licenses: MIT licenses/public domain notices
• Open Pull Request: #41386