Seq scan
Hi everybody.
Im trying to execute a simple UPDATE query on a table with 450000 rows
using an index in the where clause.
Example:
UPDATE table SET field = null WHERE field = 12345
( NOTICE: Field has an index )
But when I use the EXPLAIN command it always tells me that Postgres is
using Seq Scan, even if I disable Seq Scan in the postgresql.conf
The PostgreSQL is running on a Pentium III 900 Mhz, 512 MB RAM, HD SCSI
18 Gb with a RedHat Linux 7.3 installed.
Anyone can help me?
Thanks in advance,
Diogo Biazus
diogo@ikono.com.br
Ikono Sistemas & Automa��o
www.ikono.com.br
Hello all,
Is there a way to have the OID included with %ROWTYPE?
If not, can the record pointed to by a ROWTYPE record in a loop be
updated in place?
If not, can the record pointed by a cursor be updated in place?
If not can we add one of those to the list of "nice to have" for the
next version?
While I am at it, why not allow assignment of row variable from another
row variable?
Thanks
JLL
Jean-Luc Lachance <jllachan@nsd.ca> writes:
Is there a way to have the OID included with %ROWTYPE?
No, because there's no guarantee that a rowtype value actually came
directly from a database row. For instance, it could be formed as
the return value of a function.
regards, tom lane
Well, I found a solution for now.
I use RECORD in stead of ROWTYPE.
RECORD does hold OID if it exists.
JLL
P.S. I am still interested about updating a record in place.
Jean-Luc Lachance wrote:
Show quoted text
Hello all,
Is there a way to have the OID included with %ROWTYPE?
If not, can the record pointed to by a ROWTYPE record in a loop be
updated in place?
If not, can the record pointed by a cursor be updated in place?If not can we add one of those to the list of "nice to have" for the
next version?While I am at it, why not allow assignment of row variable from another
row variable?Thanks
JLL
---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?
I claimed victory to quickly.
PLPGSQL did compile, run and did not complain about OID but it was
always set to 0.
JLL
Jean-Luc Lachance wrote:
Show quoted text
Well, I found a solution for now.
I use RECORD in stead of ROWTYPE.
RECORD does hold OID if it exists.JLL
P.S. I am still interested about updating a record in place.
Jean-Luc Lachance wrote:
Hello all,
Is there a way to have the OID included with %ROWTYPE?
If not, can the record pointed to by a ROWTYPE record in a loop be
updated in place?
If not, can the record pointed by a cursor be updated in place?If not can we add one of those to the list of "nice to have" for the
next version?While I am at it, why not allow assignment of row variable from another
row variable?Thanks
JLL
---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere" to majordomo@postgresql.org)