include metadata from SAS files in read_sas

[rkb965]

Hello! Would it be possible to include the metadata from SAS files in read_sas?

proc contents in SAS returns information about the given dataset and its variables. AFAIK, this dataset metadata is dropped when reading into R. I would personally appreciate retaining the Created and Last Modified information, if possible.

From the proc contents documentation:

Thank you for considering this!

[jthomasmock]

Just wanted to add on a vote to this feature! The underlying C code has a header/metadata only option which has been enabled in pyreadstat already and is mentioned as an enhancement over haven: https://github.com/Roche/pyreadstat#reading-only-the-headers

I'm interested in this feature as metadata contains row/column dimensions and for very large datasets can help with say, determining how many rows to iterate across for batch processing of a larger than memory SAS file into a more open format chunk by chunk.

[muschellij2]

Old thread but you can likely do this in a lightweigth way using pandas and reticulate:

Sys.setenv("RETICULATE_PYTHON" = "managed")
library(reticulate)
py_require("pyreadstat")
pr = import("pyreadstat")
py_require("pandas")
pd = import("pandas")
chunk_size = 10000
out = pd$read_sas(input_file, chunksize=chunk_size, format="sas7bdat")
out