ArcGIS9.2+中的空间域

ArcGIS旧版本中创建要素类或要素数据集时，可以手工指定数据的空间域。但9.x中只要选择了空间参考、指定了容限值，空间域会自动确定。关于这个问题，这个解释比较全面：

 

If you set the coordinate system when creating a new feature class ordataset, we use the coordinate system's (map projection) valid extent.This is larger than the coordinate system's valid area of use. Forexample, in your case, the extent of North Carolina is the valid areaof use. The 'valid extent' is what's supported mathematically by theLambert conformal conic projection which covers a much larger area.

For geographic coordinate systems, the valid extent is greater than the world.

Weusually set the minimum extent/domain and resolution values. That oftenmeans that the maximum values are much larger than required. This couldhappen prior to 9.2 too, but wasn't as obvious.

There areseveral reasons we do this. One is that we have the informationavailable about the valid extent available now. Two, with "highprecision" aka 64bit/53bit, we can support the valid extent and alsomaintain reasonable resolution values. If you use the actual area ofuse of your data in a high precision gdb, the resolution value will beridiculously small, which can cause performance problems.

Youmay have seen the 2147483647 value before in connection with thepre-9.2 geodatabase? Values, after an internal conversion using thedomain and precision (1/resolution), had to fit between 0 and2147483647. At 9.2, we now use 2^53-2 or 9007199254740990 for the maxpermitted value. We have a lot more space to work with.

There's going to be a white paper on this, in a month or so.

So, the things to remember at 9.2 when setting up a feature class/dataset are:

1. set the coordinate system
2. The tolerance values should reflect the accuracy of your data.
2.The resolution values should be at least 10 times smaller than thecorresponding tolerance values. The resolution values dictate theminimum distance between two points in the feature class.

 

    大概意思是说：由于数据的存储精度已从25位提升到了53位（注：ArcGIS中数据均按整数存储，主要考率效率），可表达的数据范围已经足够大了，没必要再设定空间域（反正空着也没用）。所以只需注意三点：

    1. 一定要设定空间参考

    2. 容限值要能反映数据的精度

    3. 解析度必须是小于容限值的1/10（这是默认值）。