Closed randomgambit closed 6 years ago
@Ironholds how are you? What do you think about this? this is an important issue because only legacy dat
files are available before 2015 or so...
Thanks!!
I'm fine, albeit somewhat bemused as to why a 3-day delay in response was sufficient for an additional poke? What in the vignette suggests that this should work with the legacy datasets?
well you mention that
rgeolocate aims to be a single generalised package for geolocation
so one could think that it supports the legacy maxmind
dat files that were available only a few years ago.
Fair! But 'generalised' does not mean 'describes everything'; while there is a legacy C API it is not integrated here because (1) it's leaky as all heck and would make the package unstable; (2) it is rather large and the package is already somewhat above the CRAN guidelines for package size, and; (3) the legacy API has been 'legacy' since at least 2014 or, to put it another way, for 66% of MaxMind's existence as a company.
If I had the leaky temp version I'd made with the old API around I'd happily hand it over (although again, very unstable) but I'm unsure what exactly the use case is here? If you really really need to use the legacy file format for whatever reason I would recommend looking at the pygeoip module and using reticulate
to build a little R/Python binding.
@Ironholds thanks, I understand. the issue is that I cannot use python because I am using R on Spark... so if I understant there is NO R package that is able to work with the old .dat file?
Not to my knowledge; there was an extremely buggy internal prototype but absent a tarball existing in someone's email account I'd consider it dead :(
Hello @Ironholds sorry to create so many tickets right now! it is just that I absolutely need to use R to read the
maxmind
dataset and your package seem to be the only viable option!How can I use
rgeolocate
to load the legacycsv
ordat
databases (https://dev.maxmind.com/geoip/legacy/geolite/)?The vignette seem to suggest that the package than work with that, but I keep getting some errors while doing so.
Thanks!!