My solution is a new approch, hope that it’s good for you.

I use, XML feature from postgresql, XSLT processing and SVG for graphics rendering.

You can see a result with a SVG compatible browser (Firefox 3.6 for example) here : http://tmafg.free.fr/bordel/hornet/postgresql/result.xml

The idea is to generate an XML compatible with my XSLT file, so the request is :
SELECT query_to_xml(’SELECT * FROM dlt_stats WHERE dlt_id=”DLT7000-002” AND errors0 ORDER BY backup_date;’, false, false, ”);

where you have to concatenate with “”

Process the xml generated with the xsl file located here (http://tmafg.free.fr/bordel/hornet/postgresql/result.xsl) and voila !!!

My initial goal was something like this:

select dlt_id, backup_date, errors from dlt_stats order by dlt_id, backup_date;

Then plot each dlt_id on it’s own graph. This would show any trends over time.

It would be interesting to see your data set plotted against each other.

The min/max and avg is interesting. Please ignore any zero values. They will be invalid (the batch job failed to record the actual errors).

e.g:

select dlt_id, count(*), min(backup_date), max(backup_date), min(errors), max(errors), avg(errors)
from dlt_stats where errors != 0
group by dlt_id order by 7, 6 desc, 5

Thanks.

dlt=# select dlt_id, count(*), min(backup_date), max(backup_date), min(errors), max(errors), avg(errors)
from dlt_stats
group by dlt_id order by 7, 6 desc, 5;

Then I’d have a look at http://matplotlib.sourceforge.net/, there are some recipes to be using it in a web application. There’s a default interactive rendering that could be easier to use from the command line, though.