Export dataframe to SAS file with extension .sas7bdat

[ramgouveia]

I'm trying to export a R dataframe to a SAS file with extension .sas7bdat

Using write_sas() from haven package I get an error.

Is there any alternative solution?

[ttrodrigz]

Curious, what error are you getting? You're using write_sas and not write_SAS, correct?

[ramgouveia]

Yes, write_sas

Sorry for my error.

[ttrodrigz]

What error message did you receive?

[ramgouveia]

I've received no error message.

When I tried to open the dataset in SAS it just didn't opened.

[ttrodrigz]

Well, you can alternatively use the foreign package, though I have never had an issue writing a sas7bdat file with haven. Few things to check:

1. Make sure you have valid column names for SAS data sets (e.g., no periods "." in the column names). Here is a list of the rules regarding names.
2. After you write the file with write_sas(), try reading it back into R with read_sas(), if that works then it wrote correctly.
3. How are you trying to open the file in SAS? Have you tried reading it in with PROC IMPORT?

[StatSteph]

FWIW, I'm an R user and SAS user and I've given up on the write_sas in haven a long time ago. read_sas works but I just write to csv if I need to use something in SAS from R. Something as simple as this doesn't open in SAS and they are all valid variable names:

library(haven)
library(tidyverse)
mtcars %>% as_tibble(rownames = "car") %>%
   write_sas("mtcars.sas7bdat")

^{Created on 2021-11-29 by the reprex package (v2.0.1)}

SAS Universal Viewer gives an error that "Unable to load table mtcars. One common cause is when the table has an associated index which was not found".

[ttrodrigz]

Ah, good to know. It has been a while since I have needed to use SAS, but hadn't had any issues in the past when using haven and SAS Enterprise Guide.

[ramgouveia]

@StatSteph

That was the solution I've used but instead of csv, xlsx.

Thank you very much.

[nirgrahamuk]

@StatSteph

One common cause is when the table has an associated index which was not found

Makes me wonder if the rownames on the tibble is throwing it off perhaps ?

library(haven)
library(tidyverse)
mtcars %>% as_tibble(rownames = "car") %>%`row.names<-`(NULL) %>%
   write_sas("mtcars.sas7bdat")

I don't have SAS on my current system to try unfortunately.

[StatSteph]

That wouldn't change anything since tibbles don't have row.names (see below). It seems they will be removing this function from the package.

library(tidyverse)
#> Warning: package 'tibble' was built under R version 4.1.2
#> Warning: package 'readr' was built under R version 4.1.2
d1 <- mtcars %>% as_tibble(rownames = "car") 
d2 <- d1 %>%`row.names<-`(NULL)

identical(d1, d2)
#> [1] TRUE

^{Created on 2021-11-30 by the reprex package (v2.0.1)}