In WinVector/replyr: Patches to Use 'dplyr' on Remote Data Sources

Sparklyr rename fails with dev version of dplyr

Rename fails in Sparklyr, think it depends on verison of dplyr. Definitely seeing it in the dev version of dplyr as of 5-14-2017. Confirmed it works properly for dplyr 0.5.0.

Submitted as Sparklyr issue 678 and dplyr issue 2776.

# devtools::install_github("tidyverse/dplyr")
# devtools::install_github('tidyverse/dbplyr')
suppressPackageStartupMessages(library('dplyr'))
packageVersion("dplyr")
library('sparklyr')
packageVersion("sparklyr")
if(requireNamespace("dbplyr", quietly = TRUE)) {
  packageVersion("dbplyr")
}
R.Version()$version.string

dLocal <- data.frame(x = 1:2,
                     origCol = c('a', 'b'),
                     stringsAsFactors = FALSE)

sc <- sparklyr::spark_connect(version='2.0.2', 
   master = "local")

d <- copy_to(sc, dLocal, 'd')

# works
rename(dLocal, x2 = x, origCol2 = origCol)

# throws
rename(d, x2 = x, origCol2 = origCol)

spark_disconnect(sc)
rm(list=ls())
gc(verbose = FALSE)

WinVector/replyr documentation built on Oct. 22, 2020, 8:07 p.m.