Hi,
I am pondering about which datatypes should be mapped to the Torque BLOB and CLOB types, and which should not. Background is that "real" Blob and CLOB data types are handled differently by the jdbc drivers than other types, in the sense that for a CLOB or BLOB column, one does not get the value directly, but a stream from which one can read (or into which one can write). So, in a sense, for "normal" datatypes, you get the value directly, for BLOB/CLOB, you get a reference. Torque hides this different behaviour from the user.
So I would suggest that we define a minimum length for which we assume that a column is a "large object". For the columns that I have in mind to change (Postgresql BYTEA-> Blob, HSQLDB LONGVARBINARY -> BLOB, HSQLDB LONGVARCHAR -> CLOB) these objects can be up to 2^32 bytes/characters long. But if a data type in a database meets this length criterium, we do not care whether it can be accessed in a blob-scpecific way or not, we just map the BLOB/CLOB Torque types to these values.
Any comments, objections etc ?
Thomas
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
