Hello,
I have a question about ordering records based on a particular cell of a one dimensional Array.
For example, say I create a simple table (that just stores integers)...
```
gears=> CREATE TABLE foo ( value integer PRIMARY KEY);
CREATE TABLE
gears=> insert into foo Values(1);
INSERT 0 1
gears=> insert into foo Values(2);
INSERT 0 1
gears=> select * from foo;
value
-------
1
2
(2 rows)
```
Now, say I want to run a SELECT statement that also annotates each row with a an array of values, and orders the results by that new SELECTed array -- in this example we simply annotate every row with an integer array like {4,5}:
```
gears=> select *, Array[4,5] as array_from_select from foo order by array_from_select;
value | array_from_select
-------+-------------------
1 | {4,5}
2 | {4,5}
(2 rows)
```
This appears to work OK (although, of course, since every row has {4,5} annotated to it, the ORDER BY clause is pretty meaningless).
I'm confused, however, why I receive, "ERROR: column "array_from_select" does not exist" when I try to order the results by a specific index of my newly selected array. For example, when I try to order the results by the first cell of the Array, I get...
```
gears=> select *, Array[4,5] as array_from_select from foo order by array_from_select[1];
ERROR: column "array_from_select" does not exist
LINE 1: ...Array[4,5] as array_from_select from foo order by array_from...
```
Thanks for you insights and time!
- David