Neil Conway <[EMAIL PROTECTED]> writes: > (2) The spec also requires that boolean::varchar(n) should raise an > error if "n" is not large enough to accomodate the textual > representation of the boolean value.
Really? That's in direct contradiction to the "normal" spec-required behavior of casting to varchar(n). I'd suggest ignoring it on the grounds that the SQL committee have forgotten what they wrote themselves. > (3) The spec suggests that true/false should be upper-cased when > converted to text, so that's what I've implemented, but one could argue > that converting to lower-case would be more consistent with PG's general > approach to case folding. hm, +1 for lower case myself, but not dead set on it. More generally, I'm really hoping to get rid of bespoke text<->whatever cast functions in favor of using datatypes' I/O functions. To what extent can we make the boolean I/O functions serve for this? It seems relatively painless on the input side --- just allow whitespace --- but I suppose we can't change boolout's historical result of "t"/"f" without causing problems. Also, invoking btrim() seems an exceedingly expensive way of ignoring a bit of whitespace. I suppose inefficiency in a seldom-used cast function does not matter, but please don't do it that way in boolin. regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match