在redshift表中更新

Question

I have a very weird problem with a table in redshift, below are the steps that I took :

   1 : Alter table_name add columns flag integer
    /* table_name row count 10000*/
   2 :update table_name set
   flag = 1 
   where/* some condition*/
   -----rows affected 4000

3.update table_name set
flag = 0
where flag<> 1
 ----rows affected 0

When I select from table_name where flag = 1 I get 4000 then why the remaing 6000 are not getting updated? Can anyone explain logic behind this? Thanks for your help!!

Answer 1

Because flag can also be null and to compare with null you need the is operator (or a NULL save unequal operator which some DB engines provide)

update table_name 
set flag = 0
where flag is null or flag <> 1