[plt-scheme] Scsh script <-- Postgres database --> PLT program

On Tue, 2005-04-26 at 11:58 -0700, nishad at ptolemy.tlg.uci.edu wrote:
> I have a scsh script that does some logfile munging, using grep and
> wc and such like.  This script needs to get a list of IP addresses
> that are stored in a Postgres DB.
...
> I've managed (if that is the word) to get SchemeQL working on
> Solaris, so now my PLT scheme programs can access database values.
> 
> I want to write one program that does both: retrieve DB values on
> its own, and use the Unix shell tools.  Is there a way to do this,
> without the clunkiness of having the scsh program call the PLT
> program to write the DB values out to file, then read the file, and
> so on...?

There are ways.

One low-tech solution is to use the psql interface.  For the sake of
illustration, I created a table named IP_ADDRESSES and inserted three
rows into it.  You can query it by exec'ing some variation of the
following command from Scsh:

 | $ psql --host localhost --user johnsmith --command 'select * from ip_addresses' sampledb
 | Password: 
 |      ip      
 | -------------
 |  127.0.0.1
 |  192.168.0.1
 |  10.10.10.10
 | (3 rows)

If you want a more high-tech (and correspondingly more difficult to
implement) solution, you can kluge a quick and dirty partial
implementation of the PostgreSQL frontend-backend V3 protocol that
gives you just enough rope to hang yourself by sending simple one-off
SELECT queries to the backend without doing server-side prepared
statements or any of that fancier, robust stuff:

  http://www.postgresql.org/docs/current/static/protocol-flow.html#AEN55755

All you need for this approach to work is a good socket library which,
AFAIK, Scsh does have.

Both of these approaches should be doable entirely from within Scsh
without having to rely on PLT Scheme.