Search code examples
rdplyrsplittidyrlatitude-longitude

is there an R code to separate coordinates latitudes and longitudes with differing lengths?


I have included this minimal example.

cluster_id<-c(1,2)
lat_long<-c("35.92,0.34;35.98,-0.13;35.73,-1.29","38.98,-0.34;40.23,1.23")
d<-data.frame(cluster_id,lat_long)
d

I expect the following output

cluster_id<-c(1,1,1,2,2) 
latitude<-c(35.92,35.98,35.73,38.98,40.23) 
longitude<-c(0.34,-0.13,-1.29,-0.34,1.23) 
c<-data.frame(cluster_id,latitude,longitude)
c

@ Akindele Davies provided a great feedback using unsplit

However, am very interested in out put c above


Solution

  • I already answered your updated question in a comment to my original answer, but I can appreciate that it may have been hard to understand as a comment.

    First, we'll combine the steps that I laid out earlier into a function parse_line().

    parse_line <- function(line){
        coord_pairs <- strsplit(line, split = ";")
        # Separate the latitude-longitude components
        coords <- strsplit(unlist(coord_pairs), split = ",") # We have to unlist coord_pairs because strsplit() expects a character vector
        
        # coords is a list of two-element vectors (lat and long)
        # Combine the elements of coords into a matrix, then coerce to a dataframe
        
        df <- as.data.frame(do.call(rbind, coords)) 
    }
    

    Then we'll use parse_line() as a building block for a similar function parse_lines().

    parse_lines <- function(cluster_ids, lines){
      parsed_dfs <- Map(function(x, y) cbind(x, parse_line(y)), cluster_ids, lines) 
    # Iterates over all the pairs of cluster_ids and lines
    # and adds the cluster_id as a column to the dataframe produced by calling 
    # parse_line() on the corresponding line
      combined_df <- do.call(rbind, parsed_dfs) # Combines the list of dataframes into a single dataframe
      colnames(combined_df) <- c("Cluster_ID", "Latitude", "Longitude") # Adds appropriate column names
      return(combined_df)
    }
    
    parse_lines(cluster_ids, lat_long)