Re: [GENERAL] Simple, but VERYuseful enhancement for psql command - or am I missing something?

[Ben]

I want to thank everyone for replying - that was really great.

My specific issue sounds like it would be best addressed using variables.
Thanks hugely for the pointer on that one.

Displaying things vertically isn't going to work for me in particular; you
lose too much display area that way. I need to be able to see multiple
results. I'll keep \x in mind for later, though.

\e allows you to edit an external file; this is essentially the problem,
rather than the solution: I don't want to edit it outside. I want it in
the query buffer where I can just hit ^ a few times, edit right there, and
let fly. I do this all the time now, the problem is I have to hand-enter
the query the first time, or I can't do it at all.

What I was suggesting was the ability to load an external file into the
query buffer (the line buffer, maybe I mean) such that (a) it is NOT
executed a'la \i, but can be picked up into the current line for editing
using the ^ arrow, changed as required, then fired off. This is a very
general solution to having things around that you can boilerplate, or
simply use because you know they're close, even if that wasn't the idea
originally.

The idea of using & seems cool enough, but since you can use SET and then
embed them in the external file, that'll do about the same thing, it seems
to me.

Thanks again for all the responses. You people rock.

--Ben


---(end of broadcast)---
TIP 7: don't forget to increase your free space map settings

Re: [GENERAL] Simple, but VERYuseful enhancement for psql command - or am I

[Nick Barr]

Ben wrote:
I'm designing a fairly involved database system. As part fo the process, I
use the \i [FILE] command a great deal. I set up fairly involved queries,
sometimes simply for the purpose of shortening column names so the output
is reasonable. For example:
SELECT longname AS abbr,othername as "V" FROM table WHERE how;

...a bunch of these can result in a single-line output on the console,
which is a lot easier to deal with than a dump of the actual field names
which wraps around and makes you scroll back and forth trying to line up
the names with the values.
Now, in my case, I'm dealing with specific orders. So the WHERE clause
might be:
...WHERE zorder=104788;

Which works fine. But, I have to edit the file every time I'm working with
a different order, which is repetative and annoying, something computers
are supposed to save us from. :)
However, you can't leave it out; \i [FILE] expects the query to be
complete, ready to go to the server. As far as I can tell.
So - how about a command to read a file into the input lines withOUT
sending it yet, so that its ready to type the last part, such as:
   104788;

In other words, the file would end here:

...WHERE zorder=104788;
   ^
   |
   |
...then I could just type the number, hit enter, and off it would go.
Or even if it has to be complete, right now, you can use \i [FILE] and it
runs, but you can't edit the thing with the line review editing tools...
it shows the \i [FILE] command, not what the command read. That would work
too, even if it caused a dummy read the first time you used it.
Input, anyone?

--Ben

I am not sure about this exactly, but a workaround could be using 
temporary sequences. I use these a lot in some of my more involved DB 
setup scripts.

So for instance in the top level file you have:

---
CREATE SEQUENCE temp_zorder_num_seq;
SELECT setval('temp_zorder_num_seq', 104788);
\i Somefile.sql

DROP SEQUENCE
---
The in any \i file you can just use:

---
INSERT INTO some_table (zorder_num, ...) VALUES 
(currval('temp_zorder_num_seq'), ...);
---

All you have to change is the setval at the top of the script. Make sure 
you drop the sequences though ;-).

HTH

Nick

---(end of broadcast)---
TIP 6: Have you searched our list archives?
  http://archives.postgresql.org