Hi @edgararuiz. I would open an issue, but I am struggling to know how to know which package is the issue. I am testing the same thing but with RODBC as opposed to DBI
# 15.04 sec elapsed
tic()
sqlQuery2 <- "drop table [MY_DEV].[speed_test1]"
RODBC::sqlQuery(channel = conn, sqlQuery2)
df100 <- rnorm(n = 100) %>% as.tibble()
RODBC::sqlSave(channel = conn, dat = df100, tablename = "MY_DEV.speed_test1")
toc()
# 142.03 sec elapsed
tic()
sqlQuery2 <- "drop table [MY_DEV].[speed_test1]"
RODBC::sqlQuery(channel = conn, sqlQuery2)
df1e3 <- rnorm(n = 1e3) %>% as.tibble()
RODBC::sqlSave(channel = conn, dat = df1e3, tablename = "MY_DEV.speed_test1")
toc()
Maybe I can list what I have done to get pointed in the right direction.
SSIS Import Wizard write speeds to Azure are fast (TRUE)
Local (non-Azure) database write speeds as fast (TRUE)
DBI library write speeds to Azure are fast (FALSE)
RODBC library write speeds to Azure are fast (FALSE)
To me it almost seems like an Azure problem. It does not seem specific to library since DBI and RODBC were slow uploading to Azure, but fast uploading to local databases. It does not seem like a connectivity problem since the Import Wizard to Azure was fast.