describing update statements in ESQL/C

Hi. I'm trying to write a general purpose sql interpreter embedded in the interpreted language perl (isqlperl 1.2) When I finish it, It will be released under the GNU Public License, as is isqlperl 1.1. Of course you still need to buy ESQL/C to build it :-). Anyway mostly I'm OK, I've tried very hard to stay within the documented scope of ESQL/C, and I'm targetting initially at ISQL 4.10 OnLine and SE, but I have hit what seems to be a pretty major snag. I can $describe a select statement, and get back the sqlda with the sqlvar_struct array of data types being returned by the select. I can force all of these values to CCHAR_TYPE, with appropriate lengths, except BLOBs, for which I can allocate locators and treat specially (perl supports binary data which can be coerced to BYTE.) Similarily I can count the number of `?' characters in the select and allocate another sqlda/sqlvar_struct construct to pass in the host variables to the where-clause, all coerced to strings, when I open the select cursor. No problems with inserts either, I can $describe an insert statement, and get back an sqlda with the expected input data types, which I can coerce to strings, except for BLOBs, for which, again, I can allocate locator structs and treat specially. My problem is with updates. If I $describe an update statement into an sqlda, then even though the update has `?' placeholders, the sqlda.sqld (number of columns) is zero. I can of course count the `?' characters and allocate my own sqlvar_struct array, of all CCHAR_TYPE data, but then I can't pass in BLOBs to an update statement. What's the trick? I need to create a properly instantiated sqlda for a parameterised update statement, in order to look for potential BLOB data in it. Otherwise I can't UPDATE BLOB data. I know the ESQL/C manual says that you will have to parse the statement in order to determine not only the number, but also the types of the parameters, but surely a library routine exists to do this for you, even if it's undocumented. Any help would be greatly appreciated. Basically, I'm stumped. -- Cheers Bill =========================================================================== | Bill Hails <bill@tardis.co.uk> | Reisner's rule of conceptual inertia: | | C.L.I. Connect Ltd. | | | 19, Quarry St., Guildford, Surrey | If you think big enough, you'll never | | GU1 3UY. Tel (UK) 0483 300 200 | have to do it. | ===========================================================================