etfdb.R

base_url = "https://etfdb.com/data_set/?tm=40284&cond={%22by_etf%22:1272}&no_null_sort=true&count_by_id=&sort=weight&order=desc&limit=15&offset="

CleanString = function( string )
{
  result = gsub( "<a.*stock/", "", string )
  gsub( "/.*", "", result )
}

WeightsFromUrl = function( json_url )
{
  json_data     = jsonlite::read_json( json_url )
  weights_data  = sapply( json_data$rows , function( x ) x$weight )
  holdings_data = sapply( json_data$rows , function( x ) CleanString( x$holding ) )
  data.frame(weights_data, holdings_data)
}

lapply( paste0( base_url, ( 0:2 ) * 15 ), WeightsFromUrl )