[STATSGRASS] v.in.ascii

[Hamish]

> > I try to import an ascii file which is, I think, very simple:
> > x,y,z
> > 276.0447173760,10.1098557284,1030.7103
...
> > 276.0428361340,10.1088463903,840.5004
> > 276.0426783175,10.1088861157,796.0822
> > etc.........
> > 
> > in total arround 869000 lines of x,y,z values.
> > I tried this:
> > v.in.ascii input=theinputfile.txt output=theoutput fs=',' -z z=3
> > but didn't work. I've got the message:
> > Maximum input row length: 39
> > Maximum number of columns: 3
> > Minimum number of columns: 3
> > ERROR: x column is not of number type
> > That's not verbose enought for me to find what's wrong.
> > Does anybody have an idear ?
> > Thanks
> 
> I ad such a problem once, there was a space at the end of the line
> somewhere.
> You may filter the data like this:
> 
> cat theinputfile.txt | sed 's+ ++g' > theinputfile_clean.txt

fyi & iirc, G_strip() isn't used on the input lines as the first or last
column may be text, in which case the leading/trailing whitespace may be

significant. "%10s"


> then import 
> v.in.ascii input=theinputfile_clean.txt output=theoutput fs=',' -z z=3
> 
> In fact the message should indicate the offending line, but I
> don't know how to implement that. Maybe the author of v.in.ascii
> knows.
> 
> Please let us know if this trick helped.


another thing to watch: make sure there is not a blank line at the end 
of the file.

You could turn on debugging:

g.gisenv set="DEBUG=5"

Which (after a while) will show you where it broke.


> > ERROR: x column is not of number type
..
> In fact the message should indicate the offending line, but I
> don't know how to implement that.

The column is scanned, and if it contains non-numeric data anywhere it
is flagged for varchar(). It is only an error on that line if you set
column type expicitly to double or int with the columns= option*. I
guess x= y= z= should imply that.

[*] while this probably should be the case, I am not sure if it triggers
an error currently, probably not.


Hamish