On Friday, 19 August 2011, Erik Hesselink wrote:
Do you really need the precision info about the column, or do you just need the values at the right precision? Because you get the last thing already:
Prelude Database.HDBC.PostgreSQL Database.HDBC Data.Ratio Control.Monad> (fromSql . head . head) `liftM` (quickQuery db "select 1.231 ::numeric(5,0);" []) :: IO Rational 1 % 1 Prelude Database.HDBC.PostgreSQL Database.HDBC Data.Ratio Control.Monad> (fromSql . head . head) `liftM` (quickQuery db "select 1.231 ::numeric(5,4);" []) :: IO Rational 1231 % 1000
I'm not sure I understand the distinction --- to my way of thinking, getting the value at the right precision means getting the correct number of significant decimal digits, which both your example and mine fail to provide. Prelude Database.HDBC.PostgreSQL Database.HDBC Data.Ratio Control.Monad> (fromSql . head . head) `liftM` (quickQuery db "select 1.231 ::numeric(10,4);" []) :: IO Rational -- gives 1231 % 1000 == 1.231 in decimal notation Prelude Database.HDBC.PostgreSQL Database.HDBC Data.Ratio Control.Monad> (fromSql . head . head) `liftM` (quickQuery db "select 1.231 ::numeric(10,8);" []) :: IO Rational -- still gives 1231 % 1000 but should be 1.21310000 in decimal notation -- or 1231000 % 1000000 in rational notation
If you need the precision information, perhaps 'describeResult' will give you what you need. I've never used it, but it looks like it might.
Thanks for the suggestion. That will work, but ideally i'd like to have the normal usage of quickQuery, etc return correct data. Unless I am missing something here, the method of getting precision information using describeResult would be to fetch the result set, then find any columns having a numeric(x,y) SQL type and transform them. Even just getting the raw string response from the RDBMS would be better than that since the raw response is actually already a correctly formatted string representation of the data so I could use it for output without any further transformation. Is there a way to get raw response data from the RDBMS through HDBC? -- Henry House +1 530 848-1238