[GRASS-user] d.vect.thematic in wxgui

Shane Litherland litherland-farm at bigpond.com
Mon Dec 26 00:46:57 EST 2011


Hi again,

here's some of my experience with the d.vect.thematic in 6.4.2RC2 wxgui:

I tried doing a basic process through the dialog i.e. select a map,
feature type=area, pick a column, 5 intervals, most of other settings
were default.

I ran it, and it showed the vector in thematic colouring, as a new layer
in the existing display. It was listed as a layer in the layer manager
too.
Right-clicking on it in the manager and choosing properties, brought up
the d.vect.thematic dialog again, not the d.vect properties dialog that
other 'normal' vector layers showed. Seemed right to me.

Trying to use the query tool on the layer in the display window
repeatedly gave me a message that:
"no map layer selected for querying"
I checked it was the active layer in manager, and tried several places
on the thematic as I could see it in the display, with same result.

I tried zooming in/out to see if I could select a spot better (maybe I
wasn't getting the mouse click over the middle of the area, or over
where the centroid should be?..) and had a strange result.

The display did zoom, but the thematic was redrawn at the new scale
whilst the previous draw remained in place. even using the
display/render map tools in the display window did not 'clear' it up. So
after three or four zooms, I had a very colourful tessellation of my
thematic vector layer at multiple scales, but it was not very useful for
data analysis :-}

the panning tool at first seemed to be dragging the thematics and
background around in the display window, but upon releasing the mouse,
the background was redrawn in the new position but the thematics
'snapped' back to the same position they were relative to the display
window. i.e. if I had some thematic areas coloured lower-left of screen,
and I panned the display so they and background seemed to be centred,
the background would redraw with my new 'centre' but the thematic that
was lower-left still showed in lower-left. So it was no longer over the
same bit of ground, in effect.

Is this tool still a 'work in progress' to be fully functional in wxgui?
Or should it be working but I have found a bug?

A coupla questions moving past the issue of whether it works or not:

Can I use SQL or v.class or mathematical combos of columns for the
'column' input (similar to use in v.class or d.thematic.area)? It seems
I have to select a column from the table, looking at the dialog box for
this tool.

In my situation, I want to use column_3 x (column_4 squared) as the
value input for the 'column' part of the command. I can easily create
this value in the table with a bit of SQL but if I can implement this
step in GRASS it would keep the table smaller and be more versatile for
use.

I wanted to use the query tool as mentioned above, to check what values
the thematic had chosen... I'd previously used a thematic tool in tcltk
in earlier GRASS versions and noted that for a vector layer with
numerous cats for each centroid, the value chosen for the thematic
corresponded to the entry with the highest cat (generally the most
recent entry). I actually wanted to display the highest value from the
column in question for all of the entries...
Does anyone know if d.vect.thematic can be manipulated to define what
values it uses for a given area/point/etc if multiple values exist for
that area/point? in some cases, maybe I do want to display the value
from the most recent event, other times I might want to see the max or
min value at that point from all the values for that area/point.

I am OK at doing bits like this in SQL, but am somewhat lost as to how
I'd cover the geospatial aspect if GRASS itself can't do it. I am
wondering if using PostGIS to put a geometry column to the table in
question might be what I should pursue? Then I could do all the
selecting/filtering in SQL feed a 'view' (as a 'table') to GRASS to do
the d.vect.thematic on?

Hope the comments above are useful; hope someone has feedback on my
latter questions :-)

-shane.







More information about the grass-user mailing list