Thread: Coalesce 2 Arrays

Are the inputs always of fixed dimensions eg. 1 by 7?

Yes, they are. 

On Tue, Jun 25, 2019 at 4:33 AM Rob Sargent <robjsargent@gmail.com> wrote:

On Jun 24, 2019, at 2:31 PM, Alex Magnum <magnum11200@gmail.com> wrote:

Hi,
I have two arrays which I need to combine based on the individual values;

i could do a coalesce for each field but was wondering if there is an easier way

array_a{a,   null,c,   d,null,f,null}  primary

array_b{null,2   ,null,4,5   ,6,null}  secondary

 

result {a,   2,   c,   d,5,   f,null)

Any advice would be appreciated

Are the inputs always of fixed dimensions eg. 1 by 7?

create or replace function tt( a1 int[], a2 int[])
returns int[] as $$
declare
        aret int[];
        asize int;
begin
     select array_length(a1,1) into asize;
     for i in 1..asize loop
          aret[i] = coalesce(a1[i], a2[i]);
     end loop;
     return aret;
end;

$$ language plpgsql;

select * from tt(array[3,null], array[null,4]);
  tt  
-------
 {3,4}
(1 row)

Yes, they are. 

On Tue, Jun 25, 2019 at 4:33 AM Rob Sargent <robjsargent@gmail.com> wrote:

On Jun 24, 2019, at 2:31 PM, Alex Magnum <magnum11200@gmail.com> wrote:

Hi,
I have two arrays which I need to combine based on the individual values;

i could do a coalesce for each field but was wondering if there is an easier way

array_a{a,   null,c,   d,null,f,null}  primary

array_b{null,2   ,null,4,5   ,6,null}  secondary

 

result {a,   2,   c,   d,5,   f,null)

Any advice would be appreciated

Are the inputs always of fixed dimensions eg. 1 by 7?

create or replace function tt( a1 int[], a2 int[])
returns int[] as $$
declare
        aret int[];
        asize int;
begin
     select array_length(a1,1) into asize;
     for i in 1..asize loop
          aret[i] = coalesce(a1[i], a2[i]);
     end loop;
     return aret;
end;

$$ language plpgsql; select * from tt(array[3,null], array[null,4]);
  tt  
-------
 {3,4}
(1 row)

On Mon, Jun 24, 2019 at 4:11 PM Rob Sargent <robjsargent@gmail.com> wrote:

On 6/24/19 4:46 PM, Alex Magnum wrote:

Yes, they are. 

On Tue, Jun 25, 2019 at 4:33 AM Rob Sargent <robjsargent@gmail.com> wrote:

On Jun 24, 2019, at 2:31 PM, Alex Magnum <magnum11200@gmail.com> wrote:

Hi,
I have two arrays which I need to combine based on the individual values;

i could do a coalesce for each field but was wondering if there is an easier way

array_a{a,   null,c,   d,null,f,null}  primary

array_b{null,2   ,null,4,5   ,6,null}  secondary

 

result {a,   2,   c,   d,5,   f,null)

Any advice would be appreciated

Are the inputs always of fixed dimensions eg. 1 by 7?

create or replace function tt( a1 int[], a2 int[])
returns int[] as $$
declare
        aret int[];
        asize int;
begin
     select array_length(a1,1) into asize;
     for i in 1..asize loop
          aret[i] = coalesce(a1[i], a2[i]);
     end loop;
     return aret;
end;

$$ language plpgsql;

select * from tt(array[3,null], array[null,4]);
  tt  
-------
 {3,4}
(1 row)

Plain SQL variant:

 

SELECT array_agg(COALESCE(a, b))

FROM (

SELECT 

unnest(ARRAY[null, 2]::int[]),

unnest(ARRAY[1,null]::int[])

) vals (a, b);

Even if they aren't the same length the above should work, I think, as extra rows for the shorter array will contribute padded nulls.

David J.

Brilliant of course.  Maybe not as easy to stick in another query

select a.name, b.name, tt(a.intarray, b.intarray) as coalesced_array from table a join table b on a.<something> = b.<something>;

Any guess at the performance differences?