[GRASS-dev] G7: r.neighbors changes data type from CELL to DCELL

[Glynn Clements]

Markus Neteler wrote:

> >> 1) "quantile=0.5" creeps into the history:
> >>
> >> r.info landuse96_28m_mode7x7
> >> ...
> >>  |    r.neighbors input="landuse96_28m" output="landuse96_28m_mode7x7" met\   |
> >>  |    hod="mode" size=3 quantile=0.5
> >>
> >> --> quantile=0.5 should not be there (even if ignored).
> >
> > Remove the line:
> >
> >     parm.quantile->answer = "0.5";
> >
> > Optionally change:
> >
> >         out->quantile = (parm.quantile->answer && parm.quantile->answers[i])
> >             ? atof(parm.quantile->answers[i])
> >             : 0;
> >
> > to default to 0.5 rather than 0.
> >
> > But given that we already a distinct "median" method, using
> > "quantile=0.5" is largely redundant.
> 
> ... mmh, not sure what to change.

Fixed in r60345.

> >> 2) The method=mode does not preserve CELL (likewise in r.series).

Note that the same issue applies to minimum, maximum, range, and
(unweighted) sum.

> This is a bit over my programming skills.
> But the broken "mode" operator is a critical issue.
> 
> Help (=svn commits) welcome,

r60346 implements the following logic:

For any aggregate, there are 2 factors affecting the output type:

1. Whether the input map is integer or floating-point.
2. Whether the weighted or unweighted version of the aggregate is used.

These combine to create 4 possibilities:

type            integer   integer   float     float
weighted        no        yes       no        yes
			  	    	      
average         float     float     float     float
median          [1]       [1]       float     float
mode            integer   integer   [2]       [2]
minimum         integer   integer   float     float
maximum         integer   integer   float     float
range           integer   integer   float     float
stddev          float     float     float     float
sum             integer   float     float     float
count           integer   float     integer   float
variance        float     float     float     float
diversity       integer   integer   integer   integer
interspersion   integer   integer   integer   integer
quart1          [1]       [1]       float     float
quart3          [1]       [1]       float     float
perc90          [1]       [1]       float     float
quantile        [1]       [1]       float     float

[1] For integer input, quantiles may produce float results from
interpolating between adjacent values.

[2] Calculating the mode of floating-point data is essentially
meaningless.

With the current aggregates, there are 5 cases:

1. Output is always float: average, variance, stddev, quantiles (with
interpolation).

2. Output is always integer: diversity, interspersion.

3. Output is integer if unweighted, float if weighted: count.

4. Output matches input: minimum, maximum, range, mode (subject to
note 2 above), quantiles (without interpolation).

5. Output is integer for integer input and unweighted aggregate,
otherwise float: sum.

-- 
Glynn Clements <glynn at gclements.plus.com>