v.to.db sinuous

[jradinger]

The option of v.to.db to calculate sinuosity (sinuous) seems to be calculated based on features (i.e. feature ids) rather than on cats (e.g. as length is calculated). Thus, only reasonable results are calculated if a cat consists of only one feature. For connected lines that consist of more than one feature (lines) it is unclear how the sinuosity index is calculated (based on last processed feature? average over all sub-lines). For connected lines that share one cat the sinuosity should probably be calculated based on a polyline of these lines sharing a common cat.

This might be also related to ​https://trac.osgeo.org/grass/ticket/3102

[mmetz]

Replying to jradinger:

The option of v.to.db to calculate sinuosity (sinuous) seems to be calculated based on features (i.e. feature ids) rather than on cats (e.g. as length is calculated). Thus, only reasonable results are calculated if a cat consists of only one feature. For connected lines that consist of more than one feature (lines) it is unclear how the sinuosity index is calculated (based on last processed feature? average over all sub-lines).

The sinuosity index is sinuosity of the last feature processed. For multiple lines with the same category, how should sinuosity be calculated? Average of all sinuosities or sum of all line lengths / sum of all endpoint distances?

Line slope is currently the sum of all line slopes, that does not seem right.

Azimuth is the azimuth of the last feature processed. Since azimuth is an angle, a simple average over all lines with the same category would not work.

For connected lines that share one cat the sinuosity should probably be calculated based on a polyline of these lines sharing a common cat.

In this case polylines would need to be built first, before running v.to.db.

[sbl]

Markus explanation answered the question-part of the ticket quite well. However, a warning message might be helpful for the users and then there is the issue with slope Markus mentioned...