using "search" orm method querying uppercase columns, produces PostgreSQL error "column TABLE.COLUMN does not exist"

[Lorenzo Battistini]

Using upper case field name causes PostgreSQL error while searching it.

For instance, if you create a new object "custom.object" with an integer field "FIELD" and try searching it with
self.pool.get('custom.object').search(cr, uid, [('FIELD', '=', 1)])
PostgreSQL will tell "ERROR: column custom_object.field does not exist".

According to PostgreSQL documentation http://www.postgresql.org/docs/8.4/static/sql-syntax-lexical.html , "unquoted names are always folded to lower case. For example, the identifiers FOO, foo, and "foo" are considered the same by PostgreSQL, but "Foo" and "FOO" are different from these three and each other"

By first analysis, I solved modifying the "__leaf_to_sql" method of "expression" class ( http://bazaar.launchpad.net/~openerp/openobject-server/trunk/annotate/2458/bin/osv/expression.py#L342 ) replacing every
%s.%s
with
"%s"."%s"

[Lorenzo Battistini (elbati)]

Here is the patch for trunk series

[Lorenzo Battistini (elbati)]

And here is the patch for 5.0 series

[xrg (xrg)]

Your analysis and patch are technically correct. However, I would challenge the need to support all-uppercase field names. By convention, we tend to only use lowercase (not even mixed).

[Olivier Dony (Odoo) (odo-openerp)]

Thanks for the suggestion, we can keep this wishlist improvement for after v6.0.

[Olivier Dony (Odoo) (odo-openerp)]

The latest trunk contains a fix for this bug, so we can close this bug now (a patch was needed in the read() method too, in addition to expression.py).

Thanks for reporting this and for providing patches!

PS: I'm afraid such an nice-to-have improvement has very low chances to be merged in any stable version (5.0 or 6.0), as it would be against all stable policies.

[Olivier Dony (Odoo) (odo-openerp)]

For the record, the relevant trunk server revision is r.3619 revid: <email address hidden>