At present, test_databases, when encountering a failed connection, failed table creation, or other R error, will die and kill all output.
For instance,
test <- test_databases("conn.yml")
#...F....F....F....F....FF...F....F.F.EEEEEE
#Created a temporary table named: ##xkokevxfncvsbkrdejms
#....E....E....E....E....E....E....E..E...E.
# Error in connection_begin(conn@ptr) :
# nanodbc/nanodbc.cpp:1128: HYC00: [RStudio][ODBC] (11470) Transactions are not supported.
test
#Error: object 'test' not found
It is probably preferable that we either bypass the failed connection or failed table creation (and maybe propagate the error somehow), try again after waiting a little while, or both.
What I am really thinking of here is how do we best integrate with testthat? I.e. should we surface anything in the output test object about a failed connection?
At present,
test_databases, when encountering a failed connection, failed table creation, or other R error, will die and kill all output.For instance,
It is probably preferable that we either bypass the failed connection or failed table creation (and maybe propagate the error somehow), try again after waiting a little while, or both.
What I am really thinking of here is how do we best integrate with
testthat? I.e. should we surface anything in the output test object about a failed connection?