On Tue, Oct 6, 2015 at 10:04 AM, John McKown <john.archie.mck...@gmail.com> wrote:
> I'm wanting to do some reporting on data which I have an a PostgreSQL > table. For lack of anything better, I've decided to see if I can do it in > GNU awk. OK, using Perl with DBI might be a better idea, I'll grant you > that. Or maybe Python or Ruby (which I don't know). But out of shear > cussedness, I'm going to see what I can do in gawk. What I don't see is a > way to get the data out of PostgreSQL and into my awk program. Does anybody > know of a way to do this, short of "cheating" by using psql? Yes, I know > that I could do something like (BASH on Linux/Fedora 22 x86_64): > > awk -f some-program.awk <(echo 'SELECT a, b, c, d FROM schema.table ORDER > BY a' | psql dbname) > While not an expert on awk I believe this ^ is how the program is designed to function. > What I was hoping for was a "Dynamic Extension" (dll) which would allow > "native" use of PostgreSQL, or may ODBC. But I haven't found any. Have I > missed anything? If I were to create such a beastie, would it be of any use > to others? I guess, in this, I'm wondering what "report writer" most are > using when psql just doesn't have sufficient capability. > > Birt, Jasper Reports, or Crystal Reports to name a few well-known ones. > Yeah, sometimes I'm just plain stupid & pig headed. But then it will be a > "leaning experience" (with associated scars, I'm sure). > If you don't understand the nature of awk or *NIX in general (dll is a Window's thing) you are likely going to end up with much worse than scars. Amputation is likely to be required. To a large degree one is expected to use BASH (or another shell scripting language) if they want to compose programs from *NIX commands. To choose to use awk but then discard psql makes little sense. If you need something more holistic and/or capable you should just upgrade to a general-purpose language (e.g., Perl) or a custom-built solution (e.g. Birt). David J.
