I do the same, but I have one thing I miss - an easy way to send the output out for graphs.
Tweaking and playing with gnuplot is a loss of time - if on a copy/paste excel and others can understand the data from the label and plot using reasonable defaults without many hints, certainly if columns are identified as datetime, labels etc. there could be a tool to use such hints and make a decent graph (to me, decent means giving a global understanding - sure you can tweak it to look good if you are preparing a report, but a lot of time is spent graphing thinks to figure things out and many graphs go to the trash in the process)
My dream is to do my select queries in psql and direct the output to that tool, never leaving psql - so it could be for example something that would be triggered on a new table creation matching a specific name like xx_, then it would simply require prefixing "select" by "create table xx_abc as ".
The best way I've found is to save the output to a CSV and pass it to other tools, but there are never quite user friendly and usually can't pick reasonable defaults.
There is an OSX psql frontend I tried after it was recommended here on HN (http://inductionapp.com/) but it was not that helpful in day to day operations.
You could configure the PAGER environment variable and "\pset pager always" so that all your query results are passed through an external script. The external script would have to decide whether to just spit it out on stdout to display as usual, or to send the query output to a graphing program. This is a pretty ugly solution but I'd be interested to know if anyone is doing anything like this.
We have graphs in the plan for CartoDB development, and it is built on postgresql, so hopefully making it easier for people to start getting graphs out of their databses.
Tweaking and playing with gnuplot is a loss of time - if on a copy/paste excel and others can understand the data from the label and plot using reasonable defaults without many hints, certainly if columns are identified as datetime, labels etc. there could be a tool to use such hints and make a decent graph (to me, decent means giving a global understanding - sure you can tweak it to look good if you are preparing a report, but a lot of time is spent graphing thinks to figure things out and many graphs go to the trash in the process)
My dream is to do my select queries in psql and direct the output to that tool, never leaving psql - so it could be for example something that would be triggered on a new table creation matching a specific name like xx_, then it would simply require prefixing "select" by "create table xx_abc as ".
The best way I've found is to save the output to a CSV and pass it to other tools, but there are never quite user friendly and usually can't pick reasonable defaults.
There is an OSX psql frontend I tried after it was recommended here on HN (http://inductionapp.com/) but it was not that helpful in day to day operations.
Yet it seemed to be on the same problem - see this picture https://s3.amazonaws.com/induction/induction-visualize.png