What is the best way to parse lists embedded in variables within a dataframe?
When parsing json in R (I typically use the jsonlite package), I frequently end up with data frame columns containing lists (of other lists or data frames). A trivial example of this is parsing Twitter stream data where the coordinates are returned as as a list of latitude and longitude. A more complex example (and the one I am currently wrestling with) is a JSON of doctors that parses the addresses into a list of dataframes. Here is some example data illustrating the structure (this is public data, by the way):
> str(df)
Classes ‘tbl_df’ and 'data.frame': 2 obs. of 2 variables:
$ addresses:List of 2
..$ :'data.frame': 1 obs. of 6 variables:
.. ..$ address : chr "Department of Palliative Care"
.. ..$ address_2: chr "2525 Cumberland Parkway, SE"
.. ..$ city : chr "Atlanta"
.. ..$ state : chr "GA"
.. ..$ zip : chr "30305"
.. ..$ phone : chr "4043650966"
..$ :'data.frame': 2 obs. of 6 variables:
.. ..$ address : chr "5445 Meridian Mark Road" "3619 South Fulton Avenue"
.. ..$ address_2: chr "Suite 370" ""
.. ..$ city : chr "Atlanta" "Hapeville"
.. ..$ state : chr "GA" "GA"
.. ..$ zip : chr "30342" "30354"
.. ..$ phone : chr "4047652020" "4047652020"
$ npi : chr "1497831390" "1578667986"
jsonlite has a function (flatten) for extracting embedded data frames to individual variables, but it does not work on lists.
In the Twitter example, I can extract the list items to variables in the same dataframe using a for loop:
for (i in 1:nrow(df)){
#sometimes coordinates is blank, so check
if (length(df2$coordinates.coordinates[[i]]>0)){
df2[i,"coordinates.lon"]<- df2$coordinates.coordinates[[i]][1]
df2[i,"coordinates.lat"]<- df2$coordinates.coordinates[[i]][2]
}
In the Doctor example, since each Doctor can have multiple addresses, I need to create a new dataset.
library(dplyr)
addresses = data.frame()
for (i in 1:nrow(df)){
x<-df$addresses[[i]]
#need an identifier
x$id <-df[[i,"npi"]]
addresses <-bind_rows(addresses, x)
}
While both of these examples work, they are both a) slow and b) not the "R" way of doing things (as I understand it).
So, my question is: what's a better, faster, more "R" way of extracting lists from data frame variables?
Thanks to Richard Scriven. unnest
in tidr
gave me exactly what I needed.
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.