Re: Tab completion for ALTER INDEX|TABLE ALTER COLUMN SET STATISTICS

On 2018/12/26 14:50, Tatsuro Yamada wrote:
> On 2018/12/26 14:15, Michael Paquier wrote:
>> On Wed, Dec 26, 2018 at 02:05:26PM +0900, Tatsuro Yamada wrote:
>>> Do you mean my "fix_manual_of_alter_index_v2.patch"?
>>
>> Nope.  This patch is only a proposal for the documentation.  The main
>> patch to extend psql completion so as column numbers are suggested
>> fails to apply.
> 
> I rebased the WIP patch. :)
> 
> * Following query is added to get attribute numbers of index,
>    however its result contains not only expression columns but also other columns.
> 
> * I'm not sure what should I use "%d" and first "%s" in the query, so I commented out: /* %d %s */.
>    I know this is ugly.. Do you know how to use?
> 
> +#define Query_for_list_of_attribute_numbers \
> +"SELECT attnum "\
> +"  FROM pg_catalog.pg_attribute a, "\
> +"       pg_catalog.pg_class c "\
> +" WHERE c.oid = a.attrelid "\
> +"   AND a.attnum > 0 "\
> +"   AND NOT a.attisdropped "\
> +"   /* %d %s */" \
> +"   AND a.attrelid = (select oid from pg_catalog.pg_class where relname = '%s') "\
> +"   AND pg_catalog.pg_table_is_visible(c.oid) "\
> +"order by a.attnum asc "

I modified the patch to remove unusable condition.

========
# create table hoge (a integer, b integer, c integer);
# create index ind_hoge on hoge(a, b, c, (c*1), (c*2), (c*3), (c*4), (c*5), (c*6), (c*7), (c*8), (c*9));

# \d ind_hoge
        Index "public.ind_hoge"
  Column |  Type   | Key? | Definition
--------+---------+------+------------
  a      | integer | yes  | a
  b      | integer | yes  | b
  c      | integer | yes  | c
  expr   | integer | yes  | (c * 1)
  expr1  | integer | yes  | (c * 2)
  expr2  | integer | yes  | (c * 3)
  expr3  | integer | yes  | (c * 4)
  expr4  | integer | yes  | (c * 5)
  expr5  | integer | yes  | (c * 6)
  expr6  | integer | yes  | (c * 7)
  expr7  | integer | yes  | (c * 8)
  expr8  | integer | yes  | (c * 9)
btree, for table "public.hoge"

# alter index ind_hoge alter column <tab!>
1   10  11  12  2   3   4   5   6   7   8   9

# alter index ind_hoge alter column 1 <tab!>
1   10  11  12

# alter index ind_hoge alter column 10 <tab!>
alter index ind_hoge alter COLUMN 10 SET STATISTICS
========

Thanks,
Tatsuro Yamada