I am using read.xport in the foreign package in some "mission
critical"
applications. Unfortunately there remain a few bugs resulting in
erroneous values (especially more NAs than are really there) in a few
variables. The datasets I am using are unfortunately proprietary so that
I am unable to give them to the package maintainers for testing. I know
that significant improvements have been made in read.xport over the past
year, which I very much appreciate. If the maintainers could post a change
log I could at least run periodic tests using the proprietary datasets,
comparing results to dataload and to Stat/Transfer. I could not find
previous improvements to read.xport listed in the package's source file.
Thanks for considering this.
---
Frank E Harrell Jr Professor and Chair School of Medicine
Department of Biostatistics Vanderbilt University