Implement osmdata_data_frame supporting out meta; and adiff queries
#285
Conversation
|
See also #252, which should be retained as the "original" vision of this kind of function, courtesy of @Mashin6. No XML filtering; no nodes, ways, rels; just data. So any @jmaspons Would you be able to modify this to reflect #252, and return that kind of data.frame, and yet still incorporate the functionality suggested here? |
Old fill_overpass_data return either the doc as an XML object or as a
character vector, depending on the missing or not doc parameter.
All the osmdata_* functions cast the doc to character with:
rcpp_osmdata_* (paste0 (doc))
and therefore are not affected by this change.
I think the proposal is unrelated as it affects mostly
Once #252 is implemented, it should be easy to handle the case by parsing the opq$suffix or whatever flag that the implementation end up with, but I think it's premature while there are no option for different prefixes and suffixes. I will help to modify osmdata_data_frame when needed. P.S.: reading your post, I interpret that you prefer |
Profiling shows that list2DF is much faster than data.frame(list(...))
jmaspons
changed the title
out meta; and adiff queries
Fixes to get_metadata to guess query type for empty results
|
|
jmaspons
force-pushed
the
osmdata_data.frame
branch
from
8d48e7f
to
4bf05e5
Compare
|
@jmaspons I'll get to this one asap, but may take a couple of days. |
Still much slower than osmdata_sf
cherry-pick: 7fde28a
adiff have <new> node for deleted objects, but diff have not. diff results may contain duplicated entries -> unique(df)
The inputs should come from an opq object, thus, the datetimes already have passed check_datetime()
|
@jmaspons This seems good to merge now. Did you have any other changes in mind? If not, then I'll merge straight away. I had a play around with it - this really is a very nice addition! I think i'll end up using this quite a bit. |
Works and seems that doesn't break anything (check is fine). I will add tests if you like the idea.