jbryer
commented
5 years ago This seems like a fine work around. I have not intention of removing the the getIPEDSurvey anytime soon. As long as it still works, I can keep it. If this turns out to be the only viable solution for Windows, I can remove the warning message. It is unfortunate that government agencies seem to like to distribute their data in proprietary formats. NYS Dept of Education also releases their data (only) in MS Access.
On Tue, Aug 27, 2019 at 1:53 PM mikesmith2468 notifications@github.com wrote:
With the absence of the Hmisc::mdb.get function on Windows, it appears the only supported way to connect to Access databases on Windows is via ODBC. I've managed to get this working via RODBC (without needing to create the DSNs via the Windows administrative tools) but transferring the data from the Access file into R is unworkably slow for the larger tables. I got bored of waiting after 24 hours for one of them, on a reasonably capable machine. It also has significant implementation challenges depending on the "bit-ness" of both the R install and the Access install.
Instead I have a helper function that I'm testing that uses getIPEDSSurvey() to pull down all the zipped CSVs from NCES for a given year, extracts them and imports them into R individually, and then saves them all as an Rds file. My hope is that this could take the place of the download_ipeds() function on Windows machines only: as far as I have been able to tell, the resulting Rds file looks identical to the one that comes from Hmisc::mdb.get, and so your existing functions for working with the data in the Rds should not need to be changed. However, it does rely on the continued existence of the getIPEDSSurvey() function.
Thoughts?
-m.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/jbryer/ipeds/issues/9?email_source=notifications&email_token=AAEYJD5OIZYBAAMHRN7YJWDQGVSYNA5CNFSM4IQHRFB2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HHW7SZQ, or mute the thread https://github.com/notifications/unsubscribe-auth/AAEYJD7BWOD4C2G6UTE7CPLQGVSYNANCNFSM4IQHRFBQ .
mikesmith2468
tyleransom
With the absence of the Hmisc::mdb.get function on Windows, it appears the only supported way to connect to Access databases on Windows is via ODBC. I've managed to get this working via RODBC (without needing to create the DSNs via the Windows administrative tools) but transferring the data from the Access file into R is unworkably slow for the larger tables. I got bored of waiting after 24 hours for one of them, on a reasonably capable machine. It also has significant implementation challenges depending on the "bit-ness" of both the R install and the Access install.
Instead I have a helper function that I'm testing that uses getIPEDSSurvey() to pull down all the zipped CSVs from NCES for a given year, extracts them and imports them into R individually, and then saves them all as an Rds file. My hope is that this could take the place of the download_ipeds() function on Windows machines only: as far as I have been able to tell, the resulting Rds file looks identical to the one that comes from Hmisc::mdb.get, and so your existing functions for working with the data in the Rds should not need to be changed. However, it does rely on the continued existence of the getIPEDSSurvey() function.
Thoughts?
-m.