psxy gives weird results when plotting long time series (8760 records)

Added by Bernt Viggo 5 months ago

Dear GMT enthusiasts
I am experiencing weird behavior with psxy when I plot long timeseries.
Attached is a file with 8760 records.
When I plot 4000 of these records they come up like expected.
When I plot all of the 8760 records the data is not plotted correctly. Suddenly there are some spikes in the plot.
They are not the values I have in the data (pp.txt file).

This is the code for the ok_plot.jpg
R="0/4001/0/60"
awk ' { if(NR < 4000) print NR, $1; } ' pp.txt | \
gmt psxy -R$R -JX6i/2i -P -K -Bpy10 -Bpx500 > $PSFILE

Code for bad_plot.jpg
R="0/8761/0/60"
awk ' { print NR, $1; } ' pp.txt | \
gmt psxy -R$R -JX6i/2i -P -K -Bpy10 -Bpx500 -W,red > $PSFILE

Am I doing something wrong?

Thank you.

ok_plot.jpg (6.87 KB)

bad_plot.jpg (7.99 KB)

pp.txt Magnifier (59.9 KB)


Replies (4)

RE: psxy gives weird results when plotting long time series (8760 records) - Added by Joaquim 5 months ago

Your data file has 1 column only. You need at least 2 columns to do a xy plot, so I don't even know how you actully created those plots

gmt psxy -R0/4001/0/60 -JX6i/2i -P -Bpy10 -Bpx500 pp.txt > lixo.ps
psxy [ERROR]: Mismatch between actual (1) and expected (2) fields near line 1 in file
psxy [ERROR]: Mismatch between actual (1) and expected (2) fields near line 2 in file pp.dat

RE: psxy gives weird results when plotting long time series (8760 records) - Added by Paul 5 months ago

Well, he is using awk to plot that file and supply 2 columns there.
I cannot reproduce any problems using your commands with your file (well, removing -K) - looks good to me. What GMT version are you running?

RE: psxy gives weird results when plotting long time series (8760 records) - Added by Joaquim 5 months ago

Ops, visually skipped that print NR
No problems either for me.

RE: psxy gives weird results when plotting long time series (8760 records) - Added by Bernt Viggo 5 months ago

Paul wrote:

Well, he is using awk to plot that file and supply 2 columns there.
I cannot reproduce any problems using your commands with your file (well, removing -K) - looks good to me. What GMT version are you running?

Paul, Joaquim, thanks, you are the best.

I have investigated this more.
It turns out that I am using EVINCE to view the files.
There are some screen/rendering resolution issues causing this.
When I zoom in and out the plot suddenly displays correctly.
I was not aware of such issues.
So I Guess when plotting high details, the screen/zoom, effects what you actually see. Maybe I should get a new screen.
I am also using x2go, tunneling into the Linux system, maybe some issues here as well.

I am sorry for troubling you. Thanks for all help.
First time I have experienced that what you see is NOT what you get.
It looked wrong , but was correct.

(1-4/4)