I run a regression tests from 1.0.0 i.e. the tests work with 1.0.0. Look at the generated SQLs, it obviously missing the subquery using ROWNUM (using Oracle DB).
Here are the SQL stmts: OpenJPA 1.0.1: SELECT t0.id AS c0, t0.version AS c1, t0.last_updated_by AS c2, t0.last_updated_date AS c3, t1.id AS c4, t1.version AS c5, t1.last_updated_by AS c6, t1.last_updated_date AS c7, t2.id AS c8, t2.version AS c9, t2.last_updated_by AS c10, t2.last_updated_date AS c11, t2.language AS c12, t2.name AS c13, t1.content AS c14, t3.id AS c15, t3.version AS c16, t3.last_updated_by AS c17, t3.last_updated_date AS c18, t3.name AS c19, t1.load_date AS c20, t4.id AS c21, t4.type AS c22, t1.modified AS c23, t1.review_status AS c24, t1.title AS c25, t1.url AS c26, t5.id AS c27, t5.version AS c28, t5.last_updated_by AS c29, t5.last_updated_date AS c30, t5.name AS c31, t6.id AS c32, t6.version AS c33, t6.last_updated_by AS c34, t6.last_updated_date AS c35, t6.name AS c36, t7.id AS c37, t7.version AS c38, t7.last_updated_by AS c39, t7.last_updated_date AS c40, t7.name AS c41, t0.byte_length AS c42, t0.byte_offset AS c43, t0.text AS c44 FROM fact t0, document t1, collection t2, data_source t3, mime_type t4, fact_rule t5, fact_type t6, fact_type t7 WHERE (t0.fact_type_fk = ? AND t0.document_fk = ?) AND t0.document_fk = t1.id(+) AND t0.fact_rule_fk = t5.id(+) AND t0.fact_type_fk = t7.id(+) AND t1.collection_fk = t2.id(+) AND t1.data_source_fk = t3.id(+) AND t1.mime_type_fk = t4.id(+) AND t5.fact_type_fk = t6.id(+) [params=(long) 3, (long) 2] OpenJPA 1.0.0: SELECT * FROM (SELECT r.*, ROWNUM RNUM FROM (SELECT t0.id AS c0, t0.version AS c1, t0.last_updated_by AS c2, t0.last_updated_date AS c3, t1.id AS c4, t1.version AS c5, t1.last_updated_by AS c6, t1.last_updated_date AS c7, t2.id AS c8, t2.version AS c9, t2.last_updated_by AS c10, t2.last_updated_date AS c11, t2.language AS c12, t2.name AS c13, t1.content AS c14, t3.id AS c15, t3.version AS c16, t3.last_updated_by AS c17, t3.last_updated_date AS c18, t3.name AS c19, t1.load_date AS c20, t4.id AS c21, t4.type AS c22, t1.modified AS c23, t1.review_status AS c24, t1.title AS c25, t1.url AS c26, t5.id AS c27, t5.version AS c28, t5.last_updated_by AS c29, t5.last_updated_date AS c30, t5.name AS c31, t6.id AS c32, t6.version AS c33, t6.last_updated_by AS c34, t6.last_updated_date AS c35, t6.name AS c36, t7.id AS c37, t7.version AS c38, t7.last_updated_by AS c39, t7.last_updated_date AS c40, t7.name AS c41, t0.byte_length AS c42, t0.byte_offset AS c43, t0.text AS c44 FROM fact t0, document t1, collection t2, data_source t3, mime_type t4, fact_rule t5, fact_type t6, fact_type t7 WHERE (t0.fact_type_fk = ? AND t0.document_fk = ?) AND t0.document_fk = t1.id(+) AND t0.fact_rule_fk = t5.id(+) AND t0.fact_type_fk = t7.id(+) AND t1.collection_fk = t2.id(+) AND t1.data_source_fk = t3.id(+) AND t1.mime_type_fk = t4.id(+) AND t5.fact_type_fk = t6.id(+)) r WHERE ROWNUM <= ?) WHERE RNUM > ? [params=(long) 3, (long) 2, (long) 5, (long) 2] -Frank Patrick Linskey-2 wrote: > > What SQL do you see? Is this a regression from 1.0.0, or something > that you're running into for the first time? > > -Patrick > > On Nov 12, 2007 4:46 PM, frankca <[EMAIL PROTECTED]> wrote: >> >> Programmatically pagination does NOT work with OpenJPA 1.0.1 i.e. using >> EntityManager.setFirstResult()/setMaxResults(). >> >> Sorry, I don't have time to package a test case for now. >> >> -Frank >> >> -- >> View this message in context: >> http://www.nabble.com/Programmatically-pagination-does-NOT-work-with-OpenJPA-1.0.1-tf4795034.html#a13717713 >> Sent from the OpenJPA Users mailing list archive at Nabble.com. >> >> > > > > -- > Patrick Linskey > 202 669 5907 > > -- View this message in context: http://www.nabble.com/Programmatically-pagination-does-NOT-work-with-OpenJPA-1.0.1-tf4795034.html#a13717920 Sent from the OpenJPA Users mailing list archive at Nabble.com.
