st 25. 1. 2023 v 17:22 odesílatel songjinzhou <2903807914@qq.com> napsal:
As follows, we can only repeat the for statement before we use such SQL:
begin
for i in 10..20 loop
raise notice '%', i; -- Things to do
end loop;
for i in 100 .. 200 by 10 loop
raise notice '%', i; -- Things to do
end loop;
end;
But now we can simplify it as follows:
begin
for i in 10..20, 100 .. 200 by 10 loop
raise notice '%', i; -- Things to do
end loop;
end;
Although we can only use integer iterative control here, this is just a horizontal expansion of the previous logic. Thank you very much for your reply. I am very grateful!
Unfortunately, this is not a real use case - this is not an example from the real world.
And anyway, this is already supported using generate_series() and UNION:
odyssey=> do $$ declare i int; begin for i in select generate_series (10, 20) union all select generate_series (100, 200, 10) do loop raise notice 'i=%', i; end loop; end;$$; NOTICE: i=10 NOTICE: i=11 NOTICE: i=12 NOTICE: i=13 NOTICE: i=14 NOTICE: i=15 NOTICE: i=16 NOTICE: i=17 NOTICE: i=18 NOTICE: i=19 NOTICE: i=20 NOTICE: i=100 NOTICE: i=110 NOTICE: i=120 NOTICE: i=130 NOTICE: i=140 NOTICE: i=150 NOTICE: i=160 NOTICE: i=170 NOTICE: i=180 NOTICE: i=190 NOTICE: i=200 DO odyssey=>
The existing x..y notation is just syntactic sugar for a presumably common case (although I’m dubious how often one really loops through a range of numbers — surely in a database looping through a query result is overwhelmingly dominant?); I don’t think you’ll find much support around here for adding more syntax possibilities to the loop construct.
