If the field separator is <single-quote comma single-quote> then the first and last single-quotes complete the first and last field quotes Respectively and should not be removed. You simply need to specify <comma> as your field separator and single-quote and your quote symbol. The documentation for copy explains how you do this.
I can remove the single quotes from beginning and ends of all records easily. Do the single quotes need to be removed from around all fields or just from the beginnings and ends of records? I have no comas inside of fields in the file. The file first came from emacs-orgmode then was edited to create the csv file. Thanks much for help on this problem and thanks for the link to the pgloader package I'll check that out shortly.
On Fri, Jun 12, 2015 at 11:38 AM, Jude DaShiell <jdashiel@panix.com> wrote:
?[...] and ' for beginning and end of records ?[...]
?This is the problem. You need to remove the first and last single-quote and ensure that fields do not contain embedded commas. COPY will work if those two things hold true. Embedded commas and field quoting are the main difficulties with using comma-separated format.?
COPY is efficient but it expects you to provide clean data to it. You need to fix those items externally to PostgreSQL.
I have not used it myself but you might evaluate pgloader. http://pgloader.io/
