melgenek commented on code in PR #5253: URL: https://github.com/apache/arrow-datafusion/pull/5253#discussion_r1103702783
########## datafusion/core/tests/sqllogictests/README.md: ########## @@ -105,10 +105,13 @@ query <type_string> <sort_mode> - `test_name`: Uniquely identify the test name (arrow-datafusion only) - `type_string`: A short string that specifies the number of result columns and the expected datatype of each result column. There is one character in the <type_string> for each result column. The characters codes are: - - "T" for a text result, - - "I" for an integer result, - - "R" for a floating-point result, - - "?" for any other type. + - 'B' - **B**oolean, + - 'D' - **D**atetime, + - 'I' - **I**nteger, + - 'F' - **F**loating-point results, + - 'S' - **S**string, + - 'T' - **T**imestamp, + - "?" - any other types Review Comment: > I think we should stick as close as possibly to other standards -- so therefore I suggest: > T for text > P for timestamP Sounds good. It would probably make it easier to use tests suites from other databases, if you ever want to run them. Updated to be https://github.com/apache/arrow-datafusion/pull/5253/commits/7e32de81b1c238cd5932de707f12f8f40a0c9c9b. > I also think it would be good to differentiate between F and R if possible This one is trickier. It would be hard to check Postgres/Datafusion compatibility on floating points then. Postgres usually falls back to using `Numeric` by default, which is their decimal. Whereas Datafusion prefers Floats unless otherwise specified. I could technically provide a custom validator for postgres that treats expected results the same way for Postgres. So the Postgres' validator would treat `query FR` the same way as `query RR`. If you don't think having different validators is confusing, I can do it. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
