Using a named vector for dbplyr::compute.tbl_sql() fails

[nlneas1]

As a reproducible example:

sc <- sparklyr::spark_connect(master = "local")
mtcars_spark <- dplyr::copy_to(dest = sc, df = mtcars)

name <- c(name = "test")

qry <- dplyr::sql("select mpg from mtcars")

res <- sparklyr::sdf_sql(sc = sc, sql = qry)

out <- dplyr::compute(x = res, name = name)
# Error: org.apache.spark.sql.catalyst.parser.ParseException: 
# mismatched input '`name`' expecting {'(', 'SELECT', 'FROM', 'WITH', 'VALUES', 'TABLE', 'INSERT', 'MAP', 'REDUCE'}(line 2, pos 10)
# 
# == SQL ==
# CREATE TEMPORARY VIEW 
# `test` AS `name` AS SELECT `mpg`
# ----------^^^
# FROM `sparklyr_tmp_5c748cbd_9398_4dff_8c4e_f46939c88eb8`

FWIW, this seems to have become an issue after we upgraded from dbplyr 1.4.2 to 2.1.0.

[hadley]

Minimal reprex:

library(dbplyr)
library(dplyr, warn.conflicts = FALSE)

db <- memdb_frame(x = 1:10)
compute(db, c("foo" = "bar"))
#> Error: near "`foo`": syntax error

^{Created on 2021-04-01 by the reprex package (v1.0.0)}

But could you just not supply a named vector?

[nathaneastwood]

Sure, in theory. But our particular use case uses named vectors. For now, we are just using dplyr::compute(x = res, name = unname(name))