| From: | Rich Shepard <rshepard(at)appl-ecosys(dot)com> |
|---|---|
| To: | Postgresql PDX_Users <pdxpug(at)postgresql(dot)org> |
| Subject: | Re: Adding Missing Data to a Table |
| Date: | 2011-03-30 12:47:53 |
| Message-ID: | alpine.LNX.2.00.1103300542520.19520@salmo.appl-ecosys.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pdxpug |
On Tue, 29 Mar 2011, Roderick A. Anderson wrote:
> A CAST should handle getting the data from the tmp_col back into record.
> The serial was a simple way to get the column incrementally filled.
Rod,
I need to read more about CAST. My understanding is it is used to change a
data type (e.g., casting a FLOAT to NUMERIC when selecting a decimal value
that cannot be exactly matched otherwise). I did not realize it could be
used to move values among columns.
> And it appears I got caught up in the solution and forgot the record
> column was (going to be?) coming with the data from this point on. No
> need to create the SEQUENCE. Just a UNIQUE INDEX will keep it all
> straight.
Almost 1/3 of the rows already have values in the record column. I want to
replace the NULLs with values in the other two thirds so the column can be
used as a PK. Since the non-NULL rows are data from 2008-2009 it's a valid
assumption that data from 2010 and beyond will also have a value for this
attribute.
Thanks,
Rich
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Rich Shepard | 2011-04-01 15:31:19 | Re: Adding Missing Data to a Table [SOLVED] |
| Previous Message | Roderick A. Anderson | 2011-03-30 04:43:45 | Re: Adding Missing Data to a Table |