TOP 100 PERCENT is issued for MSSQL in absence of ORDER BY, breaking support for Azure DW

[alexkyllo]

PR #277 causes the mssql backend to issue SELECT TOP 100 PERCENT in sub queries even if there is no ORDER BY clause. This breaks dbplyr for Azure Data Warehouse, which is a version of MSSQL that does not support SELECT [...] PERCENT syntax. I'm working on a PR to change this so that TOP 100 PERCENT at least only gets issued if there is actually an ORDER BY, so that ADW users can use dbplyr again.

mf <- lazy_frame(x = 1:3, con = simulate_mssql())
mf %>% mutate(x = -x) %>% mutate(x = -x) %>% sql_render()
#> <SQL> SELECT -`x` AS `x`
#> FROM (SELECT TOP 100 PERCENT -`x` AS `x`
#> FROM `df`) `dbplyr_001`

[filipwastberg]

#277 also breaks code for Parallell Data Warehouse for MSSQL which doesn't support TOP ... PERCENT

[filipwastberg]

We hade to roll back to dbplyr 1.4.0 to be able to use dplyr with our database.

If anyone wonder how: devtools::install_version("dbplyr", version = "1.4.0", repos = "http://cran.us.r-project.org")