| From: | Greg Stark <stark(at)mit(dot)edu> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | John Naylor <jcnaylor(at)gmail(dot)com>, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>, David Fetter <david(at)fetter(dot)org>, Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: WIP: a way forward on bootstrap data |
| Date: | 2018-01-14 12:15:54 |
| Message-ID: | CAM-w4HMLwTz7DO-fjdCZ6FYvR7nf=xav__zE26AOKGRN54vA3Q@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
I'm 1000% on board with replacing oid constants with symbolic names
that get substituted programmatically.
However I wonder why we're bothering inventing a new syntax that
doesn't actually do much more than present static tabular data. If
things like magic proname->prosrc behaviour are not valuable then
we're not getting much out of this perl-friendly syntax that a simpler
more standard format wouldn't get us.
So just as a straw man proposal.... What if we just replaced the data
file with a csv file that could be maintained in a spreadsheet. It
could easily be parsed by perl and we could even have perl scripts
that load the records into memory and modify them. You could even
imagine writing a postgres script that loaded the csv file into a
temporary table, did complex SQL updates or other DML, then wrote it
back out to a csv file.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Emre Hasegeli | 2018-01-14 12:20:57 | Re: [HACKERS] [PATCH] Improve geometric types |
| Previous Message | Amit Khandekar | 2018-01-14 11:57:12 | Re: [HACKERS] UPDATE of partition key |