Re: Strange delete behaviour

Gnanavel S wrote:
> post the description of the t_node and t_document tables for more
> information
>
> On 8/1/05, Renzo Kottmann <renzo(at)tzi(dot)de> wrote:
>
>>Hello,
>>
>>I have a strange delete behaviour in my postgres 8.0.3 database:
>>
>>If I try a
>>
>>delete
>>from t_node
>>where node_doc_id = XX;
>>
>>from inside a plpgsql function
>>
>>on this table:
>>
>>CREATE TABLE t_node (
>>node_global_id int4 DEFAULT nextval('seq_node') NOT NULL ,
>>node_doc_id int4 NOT NULL ,
>>node_local_id int4 NOT NULL ,
>>node_offset int4 NOT NULL ,
>>FOREIGN KEY ( node_doc_id )
>>REFERENCES t_document ( doc_id )
>>MATCH FULL ,
>>PRIMARY KEY ( node_global_id )
>>);
>>
>>CREATE UNIQUE INDEX idx_node ON t_node (
>>node_doc_id
>>node_local_id
>>);
>>
>>The deletion does not finish after several minutes and the CPU is
>>running at 100% all the time unless I stop postmaster. A select works
>>normal and gives me around 2500 rows. Does anybody has an idea why this
>>happens?
>>
>>Thanks,
>>renzo

t_node see above and in addition t_document:

CREATE TABLE t_document (
doc_id int4 DEFAULT nextval('seq_document') NOT NULL ,
doc_content_id int4,
doc_lr_id int4 NOT NULL ,
doc_url text NULL ,
doc_start int4,
doc_end int4,
doc_is_markup_aware bool NOT NULL ,
FOREIGN KEY ( doc_content_id )
REFERENCES t_doc_content ( dc_id )
MATCH FULL ,
FOREIGN KEY ( doc_lr_id )
REFERENCES t_lang_resource ( lr_id )
MATCH FULL ,
PRIMARY KEY ( doc_id )
);
CREATE UNIQUE INDEX xt_document_01 on t_document (doc_lr_id);