CREATING INDEX on column having null values

Hi,


I have a partitioned table(partitioned on date) .There are about 1 million
insertions per day. There is a column called mess_id. This column is updated
.But update query is taking huge time.When i checked , this column is not
unique, and most of the time its having  null . Say everyday out of 1
million only  20K    rows are not null for this column ;.Insertions are
happening very fast , but update is  very slow.

How to optimize  my update query?  Is it good to create index on  that
column .. i m worried about my insertions getting slow

Rgrds
Suhas



--
View this message in context:
http://postgresql.1045698.n5.nabble.com/CREATING-INDEX-on-column-having-null-values-tp5735127.html
Sent from the PostgreSQL - performance mailing list archive at Nabble.com.