dbWriteTable adds `_rescued_data` column

[vladimirobucina]

When writing a table using staging volume new column is added called _rescued_data.

I've done this with mtcars dataset.

Recreate the behaviour:

library(brickster)
library(DBI)
library(testthat)

con <- dbConnect(
  drv = DatabricksSQL(),
  warehouse_id = "warehouse_id",
  catalog = "catalog"
)

dbWriteTable(
  conn = con,
  name = Id(catalog = "<catalog>", schema = "<schema>", table = "<table>"),
  value = mtcars,
  overwrite = TRUE,
  staging_volume = "/Volumes/<catalog>/<schema>/<volume>",
  progress = TRUE
)

result <- brickster::db_uc_tables_get(
  "<catalog>",
  "<schema>",
  "<table>"
)

test_that::expect_equal(result, mtcars, ignore_attr = c("class"))

Hope this example is good enough.

I'm using brickster version 0.2.11 on MacOS.

[zacdav-db]

This is due to this.

We can set schemaEvolutionMode => 'none' as a default option that we propagate is my thinking. Thoughts?

[zacdav-db]

I've decided to make it default to none and not be settable, which aligns with original expectations.

[vladimirobucina]

now it's working, great!