Thread: 1.15 DD - order of columns
Another issue with creating SQL from the DD is that unpredictably the order of columns will change. Say you have authid, authname in your original table, your copy might be defined as or authid, authname or authname, authid in the copy. Order of columns is important in Postgresql since it can't easily be changed after table is created without delete and add back in a specific order. It appears to be important with respect to import of data from CSV, for example, where even if the header defines columns in a specific order, the import does not pick that up. Workaround is of course to open up the csv in a SS, rearrange the cols to suit, and export back to CSV, or simply drop and recreate the table using the SQL from the original definition if you have that option. -- --- Colin Beckingham 613-454-5369 http://www.it4gh.com
On Sat, 2012-04-21 at 07:32 -0400, Colin Beckingham wrote: > Another issue with creating SQL from the DD is that unpredictably the > order of columns will change. Say you have authid, authname in your > original table, your copy might be defined as or authid, authname or > authname, authid in the copy. > > Order of columns is important in Postgresql since it can't easily be > changed after table is created without delete and add back in a specific > order. > > It appears to be important with respect to import of data from CSV, for > example, where even if the header defines columns in a specific order, > the import does not pick that up. > > Workaround is of course to open up the csv in a SS, rearrange the cols > to suit, and export back to CSV, or simply drop and recreate the table > using the SQL from the original definition if you have that option. As usual, I added your comments to http://code.pgadmin.org/trac/wiki/database-designer so that I can keep track of them when I'll find the time to work on it. -- Guillaume http://blog.guillaume.lelarge.info http://www.dalibo.com