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

Check durability of dplyr::arrange through dplyr::compute.

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = " # "
)
options(width =100)

library('dplyr')
library('RPostgreSQL')
packageVersion('dplyr')
packageVersion('RPostgreSQL')
my_db <- dplyr::src_postgres(host = 'localhost',port = 5432,user = 'postgres',password = 'pg')
class(my_db)
set.seed(32525)
dz <- dplyr::copy_to(my_db,data.frame(x=runif(1000)),'dz99',overwrite=TRUE)

Notice below: no warnings in frame or runtime.

dz %>% arrange(x) %>% mutate(ccol=1) %>% mutate(rank=cumsum(ccol))  -> dz1
print(dz1)
warnings()

Notice below: warning "Warning: Windowed expression 'sum("ccol")' does not have explicit order.". Result may appear the same, but we do not seem to be able to depend on that.

dz %>% arrange(x) %>% compute() %>% mutate(ccol=1) %>% mutate(rank=cumsum(ccol))  -> dz2
print(dz2)
warnings()

Notice below: warning "Warning: Windowed expression 'sum("ccol")' does not have explicit order.". Result may appear the same, but we do not seem to be able to depend on that.

dz %>% arrange(x) %>% collapse() %>% mutate(ccol=1) %>% mutate(rank=cumsum(ccol))  -> dz3
print(dz3)
warnings()

Submitted as dplyr issue 2281.

version

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