heap_mark4update: (am)invalid tid

Started by A.Bhuvaneswaranover 22 years ago2 messagesbugs
Jump to latest
#1A.Bhuvaneswaran
bhuvan@symonds.net

Hi,

I am using 7.3.2 on redhat linux 7.3.

I am getting this error during update. My update sql is,

test_pg=# update po_header set emp_name = a.user_name from users a where
emp_code = a.user_code and emp_name != a.user_name and a.user_name is not
null;

I referred the link,
http://archives.postgresql.org/pgsql-bugs/2003-04/msg00076.php .

My answers are,

(a) yes, i have before triggers
(b) no, users table donot have multi entries

Should i require to upgrade? or, Is there any smart sql?

TIA.

regards,
bhuvaneswaran

Table definitions:
Table "public.po_header"
Column | Type | Modifiers
---------------------+--------------------------+--------------------------------
po_no | text |
seq_x | integer |
seq_y | integer |
comp_code | text |
generated_by | text |
supp_code | text |
emp_code | text |
....
emp_name | text |
...
Indexes: po_header_uniq_idx unique btree (po_no, comp_code, seq_x, seq_y),
...
Check constraints: "po_header_type" ((("type" = 'NEW'::text) OR ("type" = 'CHANGED'::text)) OR ("type" = 'CANCELLED'::text))
Triggers: RI_ConstraintTrigger_17175491,
...
tapproval_date,
tbuyer_supp_map,
texception,
tmtime_po_header,
tnames_po_header,
tpo_flags_insert,
tprocess_complete,
tref_supplier,
tstatus_count

Table "public.users"
Column | Type | Modifiers
-------------------+--------------------------+-----------------------------------------------
user_id | integer | not null default nextval('user_id_seq'::text)
login_id | text |
user_code | text |
comp_code | text |
user_name | text |
...
Indexes: users_pkey primary key btree (user_id),
...
Check constraints: "users_email_id" ((email_id ~ '@'::text) AND (email_id ~ '.'::text))
Triggers: RI_ConstraintTrigger_17175378,
...
tmtime_users

#2Tom Lane
tgl@sss.pgh.pa.us
In reply to: A.Bhuvaneswaran (#1)
Re: heap_mark4update: (am)invalid tid

"A.Bhuvaneswaran" <bhuvan@symonds.net> writes:

Should i require to upgrade? or, Is there any smart sql?

If you still see the problem after updating to 7.3.4, it would be worth
investigating further. But given that the symptom looks just like a
known 7.3.2 bug, I don't see any value in spending time on it before
you update.

regards, tom lane