<html>
<head>
<meta http-equiv="Content-Type" content="text/html;
charset=windows-1252">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Even Rouault kirjoitti 26.10.2017 klo 20:17:<br>
<blockquote type="cite" cite="mid:2224838.mgomoDzED3@even-i700">
<meta name="qrichtext" content="1">
<style type="text/css">
p, li { white-space: pre-wrap; }</style>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> Would it be possible to keep X,Y as it is and add support for</p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> multidimensional data as an extension?</p>
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">That could be indeed an option if we are sure that the 2 first dimensions are X and Y, and we want only to slice higher dimensionnal variables to X,Y.</p>
</blockquote>
<br>
Assuming we have a X,Y,Z,T dataset of sea temperatures (Z is depth),
we may want to warp it from projection A to B and then select a
location and look a Z,T image from that location. Warping would work
on X,Y and compute the Z,T as an extension what it now does for the
pixel values (possibly multiple bands). I could imagine that the
user would want to use GDAL for slicing the Z,T into an image - GDAL
can write image files, so why not?<br>
<br>
So perhaps the assumption for only slicing to X,Y is not ok.<br>
<br>
<blockquote type="cite" cite="mid:2224838.mgomoDzED3@even-i700">
<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> </p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">So instead of having a general purpose indexing mechanism as I scetched initially an option would be to expose the extra dimensions/axis, their bounds. The user could set the values at which to set the values at which those extra dimensions must be considered, so as to get a fully constrainted X,Y indexed field.</p>
<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">The main advantage of this is that it would have presumably no impact on 2D-only drivers.</p>
</blockquote>
<br>
I assume that would be very good from the point of view of actual
feasibility of the whole project.<br>
<br>
Ari<br>
<br>
</body>
</html>