14-hybrid-programming.Rmd

# Hybrid Programming with SAS and R

This section contains current solutions of hybrid programming.

## Use {sasr}

`{[sasr](https://github.com/insightsengineering/sasr)}` is a package built on
`{[reticulate](https://rstudio.github.io/reticulate/index.html)}` so that the Python package
`{[saspy](https://github.com/sassoftware/saspy)}`, an official python interface to connect SAS, can be called from R.
It adds some simple wrappers to the {saspy} functionalities to make the syntax similar to R.

Here are some simple examples of how you can use it.
```{r eval = FALSE}
library(sasr)
# Transfer data from R to SAS.
df2sd(mtcars)

# Execute SAS commands.
run_sas("
  data test;
    set work.mtcars;
  run;
")

# Transfer data back to R.
sd2df("test")
```
This requires that you have a machine with SAS and it accessible through specific
[connection types](https://sassoftware.github.io/saspy/configuration.html),
where `ssh` is a common type of connection.

## Use {jupyter}/{quarto} and {saspy}

[jupyter](http://www.jupyter.org/) and [quarto](https://quarto.org/) are both capable of creating dynamic outputs from multiple languages.
Although the official SAS kernel is not supported, it can be installed through package [sas_kernel](https://github.com/sassoftware/sas_kernel) provided by [SAS software](https://github.com/sassoftware).

A quick example can be found at [Harvey Lieberman's blog](https://www.harveyl888.com/post/2022-06-27-quarto_sas_01/).

The requirement is the same as {sasr} as they both use {saspy} as the backend.

## Use {SASmarkdown}

{[SASmarkdown](https://github.com/Hemken/SASmarkdown)} is a package that installs the SAS engine to {knitr} so that in your `.Rmd` files
using SAS engines is possible.

However, this requires that SAS is installed locally on your machine, in contrast to client to a server.

## SAS with command line or ssh (not recommended)

As SAS can be called through command line, we can actually use the following to execute SAS codes

```{sh eval = FALSE}
sas script.sas
```

or, we can call SAS in R side

```{r eval = FALSE}
system("sas script.sas")
# using ssh

library(ssh)
ssh_connect()
scp_upload()
ssh_exec("sas script.sas")
```

Please note that using command line, you will need to have SAS installed on your machine.