Matt Van Mater wrote:
> I have a comment about the following thread from about 2 months ago that
> dealt with delimiters in the copy command:
> http://marc.theaimsgroup.com/?l=postgresql-general&m=107960271207890&w=2
>
> I wanted to chime in and say that I think having more complex delimiters
> would be very useful. I was curious if there are any developers considering
> working on this? Here is an example of why I would like to see multi
> character delimiters:
>
> I use pgsql 7.3.2 on OpenBSD 3.4 to store the results of my nessus scans
> (nessus is a very large and popular vulnerability scanner, see nessus.org).
> I currently use perl to parse the scan results and create tens of thousands
> of insert statements for each scan that is run, and was looking into using
> the copy command to speed up adding that data to the db. Nessus scan
> results are very text heavy and often have many characters within the text
> fields, so I don't think there are any single characters that I can use as a
> delimiter (^|,\n and many others are all used within the scan results).
> With that in mind I'd like to specify my own custom multi character
> delimiter to ensure that it won't be used within a field.
It is unlikely we would add anything like this to pgsql because it is
not deterministic. I would think the new quote setup should work, but
it does require you to double-quotes in the text.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073